{ "$schema": "https://gentic.news/research-frontier/schema.json", "name": "gentic.news Deployment Atlas", "version": "1.0", "license": "CC-BY-4.0", "license_url": "https://creativecommons.org/licenses/by/4.0/", "generated_at": "2026-04-23T07:11:44.602Z", "citation": "gentic.news Deployment Atlas (2026). https://gentic.news/research-frontier", "homepage": "https://gentic.news/research-frontier", "methodology": "https://gentic.news/research-frontier/methodology", "stats": { "total_edges": 192, "median_velocity_days": 1496, "p90_velocity_days": 3200, "fastest_deploy": { "product_slug": "llama-4-maverick", "product_name": "Llama 4 Maverick", "technique_slug": "yarn-context-extension", "technique_name": "YaRN RoPE Context Extension", "category": "architecture", "origin_date": "2023-08-31", "deploy_date": "2025-04-05T00:00:00+00:00", "velocity_days": 583, "inventor_lab": "Nous Research" }, "slowest_deploy": { "product_slug": "kimi-k2-6", "product_name": "Kimi K2.6", "technique_slug": "mixture-of-experts", "technique_name": "Mixture of Experts (Sparse MoE for LLMs)", "category": "architecture", "origin_date": "2017-01-23", "deploy_date": "2026-04-20T00:00:00+00:00", "velocity_days": 3374, "inventor_lab": "Google" }, "year_trends": [ { "year": 2025, "median_velocity_days": 1499, "n": 9 }, { "year": 2026, "median_velocity_days": 1496, "n": 183 } ] }, "techniques": [ { "slug": "awq-quantization", "name": "AWQ (Activation-Aware Weight Quantization)", "category": "inference", "origin_date": "2023-06-01", "inventor_lab": "MIT", "description": "4-bit weight quantization that preserves salient weights based on activation magnitudes, matching GPTQ quality with faster inference.", "origin_paper_url": "https://arxiv.org/abs/2306.00978", "deployments_count": 0 }, { "slug": "clip", "name": "CLIP (Contrastive Language-Image Pretraining)", "category": "multimodal", "origin_date": "2021-02-26", "inventor_lab": "OpenAI", "description": "Dual-encoder model trained on 400M image-caption pairs to align image and text embeddings, enabling zero-shot visual classification.", "origin_paper_url": "https://arxiv.org/abs/2103.00020", "deployments_count": 1 }, { "slug": "chain-of-thought", "name": "Chain-of-Thought Prompting", "category": "reasoning", "origin_date": "2022-01-28", "inventor_lab": "Google", "description": "A prompting technique that elicits step-by-step reasoning by showing exemplars that include intermediate reasoning steps.", "origin_paper_url": "https://arxiv.org/abs/2201.11903", "deployments_count": 25 }, { "slug": "chinchilla-scaling", "name": "Chinchilla Scaling Laws", "category": "training", "origin_date": "2022-03-29", "inventor_lab": "Google DeepMind", "description": "Scaling law showing compute-optimal models use ~20 training tokens per parameter — correcting prior over-parameterization in GPT-3-era models.", "origin_paper_url": "https://arxiv.org/abs/2203.15556", "deployments_count": 0 }, { "slug": "constitutional-ai", "name": "Constitutional AI", "category": "alignment", "origin_date": "2022-12-15", "inventor_lab": "Anthropic", "description": "Training harmless assistants using a written constitution of principles and an AI-generated critique/revision loop rather than human labels for every case.", "origin_paper_url": "https://arxiv.org/abs/2212.08073", "deployments_count": 7 }, { "slug": "continuous-batching", "name": "Continuous Batching", "category": "inference", "origin_date": "2022-07-01", "inventor_lab": "Seoul National University", "description": "A scheduling technique that adds/removes requests at the iteration level rather than the batch level, dramatically increasing throughput for LLM serving.", "origin_paper_url": "https://www.usenix.org/conference/osdi22/presentation/yu", "deployments_count": 0 }, { "slug": "deep-rl-from-human-preferences", "name": "Deep RL from Human Preferences", "category": "alignment", "origin_date": "2017-06-12", "inventor_lab": "OpenAI", "description": "Learning reward functions from pairwise human comparisons rather than hand-coded rewards. The direct precursor to RLHF.", "origin_paper_url": "https://arxiv.org/abs/1706.03741", "deployments_count": 0 }, { "slug": "dense-retrieval-dpr", "name": "Dense Passage Retrieval (DPR)", "category": "retrieval", "origin_date": "2020-04-10", "inventor_lab": "Meta AI", "description": "Learned dual-encoder retrieval that outperforms BM25 on open-domain QA by training encoders on question-passage pairs.", "origin_paper_url": "https://arxiv.org/abs/2004.04906", "deployments_count": 0 }, { "slug": "dpo", "name": "Direct Preference Optimization (DPO)", "category": "alignment", "origin_date": "2023-05-29", "inventor_lab": "Stanford", "description": "Aligning LMs to preference data by directly optimizing a closed-form likelihood ratio, eliminating the reward model and RL loop of RLHF.", "origin_paper_url": "https://arxiv.org/abs/2305.18290", "deployments_count": 0 }, { "slug": "flamingo-vlm", "name": "Flamingo (Cross-Attention VLMs)", "category": "multimodal", "origin_date": "2022-04-29", "inventor_lab": "Google DeepMind", "description": "Cross-attention layers interleaved into a frozen LLM that attend to vision features, enabling few-shot visual question answering.", "origin_paper_url": "https://arxiv.org/abs/2204.14198", "deployments_count": 0 }, { "slug": "flash-attention", "name": "FlashAttention", "category": "inference", "origin_date": "2022-05-27", "inventor_lab": "Stanford", "description": "A tiled, IO-aware attention kernel that computes exact attention with linear memory by fusing reads/writes to SRAM.", "origin_paper_url": "https://arxiv.org/abs/2205.14135", "deployments_count": 12 }, { "slug": "gptq", "name": "GPTQ Quantization", "category": "inference", "origin_date": "2022-10-31", "inventor_lab": "ISTA", "description": "Post-training quantization to 3-4 bits using second-order information, enabling 175B-scale LLMs to run on single-GPU inference.", "origin_paper_url": "https://arxiv.org/abs/2210.17323", "deployments_count": 0 }, { "slug": "grouped-query-attention", "name": "Grouped-Query Attention (GQA)", "category": "architecture", "origin_date": "2023-05-22", "inventor_lab": "Google", "description": "An inference-time optimization that groups multiple query heads to share a single key/value head, reducing KV cache memory at minimal quality loss.", "origin_paper_url": "https://arxiv.org/abs/2305.13245", "deployments_count": 10 }, { "slug": "int8-llm-quantization", "name": "INT8 Weight Quantization for LLMs", "category": "inference", "origin_date": "2022-08-15", "inventor_lab": "University of Washington", "description": "Row-wise and vector-wise INT8 quantization with outlier detection that enables zero-degradation 8-bit inference of LLMs.", "origin_paper_url": "https://arxiv.org/abs/2208.07339", "deployments_count": 0 }, { "slug": "ipo", "name": "Identity Preference Optimization (IPO)", "category": "alignment", "origin_date": "2023-10-18", "inventor_lab": "Google DeepMind", "description": "A preference-optimization variant that avoids DPO's over-fitting by adding an explicit regularizer.", "origin_paper_url": "https://arxiv.org/abs/2310.12036", "deployments_count": 0 }, { "slug": "instruction-tuning-flan", "name": "Instruction Tuning (FLAN)", "category": "training", "origin_date": "2021-09-03", "inventor_lab": "Google", "description": "Fine-tuning a pretrained LM on a mixture of tasks phrased as natural-language instructions, enabling strong zero-shot generalization.", "origin_paper_url": "https://arxiv.org/abs/2109.01652", "deployments_count": 17 }, { "slug": "kto", "name": "KTO (Kahneman-Tversky Optimization)", "category": "alignment", "origin_date": "2024-02-02", "inventor_lab": "Contextual AI", "description": "Alignment method that treats individual completions as binary-good/bad signals (no preference pairs needed) inspired by prospect theory.", "origin_paper_url": "https://arxiv.org/abs/2402.01306", "deployments_count": 0 }, { "slug": "llava-vit", "name": "LLaVA (Visual Instruction Tuning)", "category": "multimodal", "origin_date": "2023-04-17", "inventor_lab": "University of Wisconsin", "description": "Projecting CLIP features into an LLM's token space via a simple projector + instruction tuning on GPT-4-generated visual conversations.", "origin_paper_url": "https://arxiv.org/abs/2304.08485", "deployments_count": 3 }, { "slug": "latent-diffusion", "name": "Latent Diffusion", "category": "multimodal", "origin_date": "2021-12-20", "inventor_lab": "LMU Munich / RunwayML", "description": "Diffusion performed in a compressed VAE latent space, making high-resolution image generation tractable on consumer GPUs.", "origin_paper_url": "https://arxiv.org/abs/2112.10752", "deployments_count": 0 }, { "slug": "lora", "name": "LoRA (Low-Rank Adaptation)", "category": "training", "origin_date": "2021-06-17", "inventor_lab": "Microsoft", "description": "Parameter-efficient fine-tuning that injects low-rank decomposition matrices into attention weights, training <1% of parameters.", "origin_paper_url": "https://arxiv.org/abs/2106.09685", "deployments_count": 1 }, { "slug": "mamba-ssm", "name": "Mamba / Selective State Space Models", "category": "architecture", "origin_date": "2023-12-01", "inventor_lab": "CMU", "description": "A state-space sequence model with input-dependent selection that matches Transformer quality with linear inference cost and unlimited context.", "origin_paper_url": "https://arxiv.org/abs/2312.00752", "deployments_count": 1 }, { "slug": "mixture-of-depths", "name": "Mixture of Depths", "category": "architecture", "origin_date": "2024-04-02", "inventor_lab": "Google DeepMind", "description": "A technique letting tokens skip transformer layers when unnecessary, allocating compute adaptively based on token importance.", "origin_paper_url": "https://arxiv.org/abs/2404.02258", "deployments_count": 0 }, { "slug": "mixture-of-experts", "name": "Mixture of Experts (Sparse MoE for LLMs)", "category": "architecture", "origin_date": "2017-01-23", "inventor_lab": "Google", "description": "An architecture where a router activates only a subset of expert sub-networks per token, scaling parameter count without proportional compute cost.", "origin_paper_url": "https://arxiv.org/abs/1701.06538", "deployments_count": 17 }, { "slug": "paged-attention", "name": "PagedAttention (vLLM)", "category": "inference", "origin_date": "2023-09-12", "inventor_lab": "UC Berkeley", "description": "A memory-management scheme for KV cache modeled on OS paging, eliminating fragmentation and enabling high-throughput serving.", "origin_paper_url": "https://arxiv.org/abs/2309.06180", "deployments_count": 1 }, { "slug": "process-reward-models", "name": "Process Reward Models", "category": "reasoning", "origin_date": "2023-05-31", "inventor_lab": "OpenAI", "description": "Reward models trained to score each intermediate reasoning step rather than only the final answer, enabling superior reasoning policy learning.", "origin_paper_url": "https://arxiv.org/abs/2305.20050", "deployments_count": 1 }, { "slug": "qlora", "name": "QLoRA", "category": "training", "origin_date": "2023-05-23", "inventor_lab": "University of Washington", "description": "LoRA fine-tuning on 4-bit quantized base weights, enabling 65B-model fine-tuning on a single 48GB GPU.", "origin_paper_url": "https://arxiv.org/abs/2305.14314", "deployments_count": 0 }, { "slug": "rlaif", "name": "RLAIF (Reinforcement Learning from AI Feedback)", "category": "alignment", "origin_date": "2023-09-01", "inventor_lab": "Google", "description": "Using an off-the-shelf LLM to generate preference labels, scaling preference learning without human annotators.", "origin_paper_url": "https://arxiv.org/abs/2309.00267", "deployments_count": 0 }, { "slug": "react-agent", "name": "ReAct (Reason + Act)", "category": "agents", "origin_date": "2022-10-06", "inventor_lab": "Princeton / Google", "description": "An agent pattern that interleaves reasoning traces with tool-use actions, using each observation to refine the next reasoning step.", "origin_paper_url": "https://arxiv.org/abs/2210.03629", "deployments_count": 3 }, { "slug": "red-teaming-preference", "name": "Red-Teaming with Preference Models", "category": "alignment", "origin_date": "2022-02-07", "inventor_lab": "Google DeepMind", "description": "Using an LM to generate adversarial prompts that elicit harmful behavior, scaling safety evaluation far beyond human red-teaming.", "origin_paper_url": "https://arxiv.org/abs/2202.03286", "deployments_count": 0 }, { "slug": "reflexion", "name": "Reflexion", "category": "agents", "origin_date": "2023-03-20", "inventor_lab": "Northeastern / MIT", "description": "Agent framework that converts environment feedback into verbal self-reflection stored in memory, improving performance across trials without weight updates.", "origin_paper_url": "https://arxiv.org/abs/2303.11366", "deployments_count": 0 }, { "slug": "rlhf", "name": "Reinforcement Learning from Human Feedback (RLHF)", "category": "alignment", "origin_date": "2022-03-04", "inventor_lab": "OpenAI", "description": "A three-stage recipe (SFT → reward model from human comparisons → PPO) that aligns LM outputs with human preferences. InstructGPT is the canonical reference.", "origin_paper_url": "https://arxiv.org/abs/2203.02155", "deployments_count": 3 }, { "slug": "rejection-sampling-ft", "name": "Rejection Sampling Fine-Tuning", "category": "training", "origin_date": "2023-07-18", "inventor_lab": "Meta AI", "description": "Sampling multiple completions, scoring with a reward model, and fine-tuning on the top samples — a simpler alternative to PPO used in Llama 2.", "origin_paper_url": "https://arxiv.org/abs/2307.09288", "deployments_count": 0 }, { "slug": "rag", "name": "Retrieval-Augmented Generation (RAG)", "category": "retrieval", "origin_date": "2020-05-22", "inventor_lab": "Meta AI", "description": "Conditioning generation on retrieved passages from a non-parametric memory, combining parametric and retrieval-based knowledge.", "origin_paper_url": "https://arxiv.org/abs/2005.11401", "deployments_count": 1 }, { "slug": "rope-position", "name": "Rotary Position Embedding (RoPE)", "category": "architecture", "origin_date": "2021-04-20", "inventor_lab": "Zhuiyi Technology", "description": "A relative-position encoding that rotates query/key vectors in complex space, giving transformers better length extrapolation than absolute sinusoidal embeddings.", "origin_paper_url": "https://arxiv.org/abs/2104.09864", "deployments_count": 23 }, { "slug": "self-consistency", "name": "Self-Consistency", "category": "reasoning", "origin_date": "2022-03-21", "inventor_lab": "Google", "description": "Sample multiple CoT completions and take the majority-vote answer, substantially improving reasoning accuracy.", "origin_paper_url": "https://arxiv.org/abs/2203.11171", "deployments_count": 11 }, { "slug": "self-instruct", "name": "Self-Instruct", "category": "training", "origin_date": "2022-12-20", "inventor_lab": "University of Washington", "description": "Bootstrapping instruction-tuning data by having an LM generate its own instructions, inputs, and outputs from a small seed set.", "origin_paper_url": "https://arxiv.org/abs/2212.10560", "deployments_count": 0 }, { "slug": "self-rewarding-lm", "name": "Self-Rewarding Language Models", "category": "alignment", "origin_date": "2024-01-18", "inventor_lab": "Meta AI", "description": "Iterative alignment where the LM judges its own outputs using an LLM-as-a-judge prompt, removing human-labeled preferences from the loop.", "origin_paper_url": "https://arxiv.org/abs/2401.10020", "deployments_count": 0 }, { "slug": "sparse-autoencoders", "name": "Sparse Autoencoders for Interpretability", "category": "interpretability", "origin_date": "2023-10-05", "inventor_lab": "Anthropic", "description": "Training sparse autoencoders on residual-stream activations to extract monosemantic, human-interpretable features from transformer internals.", "origin_paper_url": "https://transformer-circuits.pub/2023/monosemantic-features", "deployments_count": 0 }, { "slug": "speculative-decoding", "name": "Speculative Decoding", "category": "inference", "origin_date": "2022-11-30", "inventor_lab": "Google", "description": "An inference technique where a small draft model proposes tokens and a large model verifies them in parallel, yielding 2-3x speedup without quality loss.", "origin_paper_url": "https://arxiv.org/abs/2211.17192", "deployments_count": 1 }, { "slug": "streaming-llm", "name": "StreamingLLM (Attention Sinks)", "category": "inference", "origin_date": "2023-09-29", "inventor_lab": "MIT", "description": "A sliding-window attention pattern with preserved initial tokens (\"sinks\") that enables indefinite streaming generation without quality collapse.", "origin_paper_url": "https://arxiv.org/abs/2309.17453", "deployments_count": 0 }, { "slug": "synthetic-data-orca", "name": "Synthetic Data Distillation (Orca)", "category": "training", "origin_date": "2023-06-05", "inventor_lab": "Microsoft Research", "description": "Training smaller models on GPT-4-generated explanation traces rather than answer-only demonstrations, closing the capability gap.", "origin_paper_url": "https://arxiv.org/abs/2306.02707", "deployments_count": 0 }, { "slug": "test-time-compute-scaling", "name": "Test-Time Compute Scaling", "category": "reasoning", "origin_date": "2024-08-06", "inventor_lab": "Google DeepMind", "description": "Allocating more compute at inference (longer reasoning chains, multiple samples + verifier) can outperform scaling parameters — the basis for o1-style reasoning models.", "origin_paper_url": "https://arxiv.org/abs/2408.03314", "deployments_count": 5 }, { "slug": "toolformer", "name": "Toolformer (Tool Use)", "category": "agents", "origin_date": "2023-02-09", "inventor_lab": "Meta AI", "description": "Self-supervised approach where an LM learns when and how to call external APIs by generating and filtering its own tool-use demonstrations.", "origin_paper_url": "https://arxiv.org/abs/2302.04761", "deployments_count": 0 }, { "slug": "transformer-self-attention", "name": "Transformer Self-Attention", "category": "architecture", "origin_date": "2017-06-12", "inventor_lab": "Google", "description": "A sequence-to-sequence architecture that replaces recurrence with scaled dot-product attention, enabling parallel training and long-range context modeling.", "origin_paper_url": "https://arxiv.org/abs/1706.03762", "deployments_count": 25 }, { "slug": "tree-of-thoughts", "name": "Tree of Thoughts", "category": "reasoning", "origin_date": "2023-05-17", "inventor_lab": "Princeton / Google DeepMind", "description": "Reasoning over a tree of intermediate thoughts with explicit look-ahead, backtracking, and self-evaluation, beyond linear CoT.", "origin_paper_url": "https://arxiv.org/abs/2305.10601", "deployments_count": 0 }, { "slug": "vit", "name": "Vision Transformer (ViT)", "category": "multimodal", "origin_date": "2020-10-22", "inventor_lab": "Google", "description": "Applying a standard Transformer directly to sequences of image patches, establishing Transformers as the dominant image-recognition backbone.", "origin_paper_url": "https://arxiv.org/abs/2010.11929", "deployments_count": 4 }, { "slug": "whisper-asr", "name": "Whisper (Robust Speech Recognition)", "category": "multimodal", "origin_date": "2022-12-06", "inventor_lab": "OpenAI", "description": "Encoder-decoder Transformer trained on 680k hours of weakly-supervised multilingual speech, setting new robustness benchmarks across accents and noise.", "origin_paper_url": "https://arxiv.org/abs/2212.04356", "deployments_count": 0 }, { "slug": "yarn-context-extension", "name": "YaRN RoPE Context Extension", "category": "architecture", "origin_date": "2023-08-31", "inventor_lab": "Nous Research", "description": "A method to extend RoPE-based models to much longer contexts via frequency-dependent interpolation, with minimal fine-tuning data.", "origin_paper_url": "https://arxiv.org/abs/2309.00071", "deployments_count": 9 }, { "slug": "zero-shot-cot", "name": "Zero-Shot Chain-of-Thought", "category": "reasoning", "origin_date": "2022-05-24", "inventor_lab": "University of Tokyo", "description": "Eliciting step-by-step reasoning without few-shot exemplars, simply by appending a phrase like \"let's think step by step\".", "origin_paper_url": "https://arxiv.org/abs/2205.11916", "deployments_count": 11 } ], "graph": { "nodes": [ { "id": "f41e091c-2cdb-4589-9734-ced31b18b372", "slug": "transformer-self-attention", "name": "Transformer Self-Attention", "type": "technique", "category": "architecture", "origin_date": "2017-06-12", "inventor_lab": "Google" }, { "id": "53b1c32a-4b47-4077-8c67-81060cbb6420", "slug": "rope-position", "name": "Rotary Position Embedding (RoPE)", "type": "technique", "category": "architecture", "origin_date": "2021-04-20", "inventor_lab": "Zhuiyi Technology" }, { "id": "8608614e-381c-464e-a424-8ce991c97ef4", "slug": "grouped-query-attention", "name": "Grouped-Query Attention (GQA)", "type": "technique", "category": "architecture", "origin_date": "2023-05-22", "inventor_lab": "Google" }, { "id": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "slug": "mixture-of-experts", "name": "Mixture of Experts (Sparse MoE for LLMs)", "type": "technique", "category": "architecture", "origin_date": "2017-01-23", "inventor_lab": "Google" }, { "id": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "slug": "flash-attention", "name": "FlashAttention", "type": "technique", "category": "inference", "origin_date": "2022-05-27", "inventor_lab": "Stanford" }, { "id": "30e88a7b-e2b0-4900-905a-9a82f729f4fa", "slug": "mamba-ssm", "name": "Mamba / Selective State Space Models", "type": "technique", "category": "architecture", "origin_date": "2023-12-01", "inventor_lab": "CMU" }, { "id": "393b2e97-09d8-45bd-ba1a-65c84065e17e", "slug": "mixture-of-depths", "name": "Mixture of Depths", "type": "technique", "category": "architecture", "origin_date": "2024-04-02", "inventor_lab": "Google DeepMind" }, { "id": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "slug": "yarn-context-extension", "name": "YaRN RoPE Context Extension", "type": "technique", "category": "architecture", "origin_date": "2023-08-31", "inventor_lab": "Nous Research" }, { "id": "f1d61976-639f-4df6-b2d2-7bc16dbaa86b", "slug": "speculative-decoding", "name": "Speculative Decoding", "type": "technique", "category": "inference", "origin_date": "2022-11-30", "inventor_lab": "Google" }, { "id": "91fa092e-f841-4016-93a8-7f2716090cc0", "slug": "paged-attention", "name": "PagedAttention (vLLM)", "type": "technique", "category": "inference", "origin_date": "2023-09-12", "inventor_lab": "UC Berkeley" }, { "id": "1acc4d7b-22bd-4d6f-a4f0-cb45d8a0eb77", "slug": "continuous-batching", "name": "Continuous Batching", "type": "technique", "category": "inference", "origin_date": "2022-07-01", "inventor_lab": "Seoul National University" }, { "id": "efa8731f-4888-4f0a-97d5-60744bf5a113", "slug": "streaming-llm", "name": "StreamingLLM (Attention Sinks)", "type": "technique", "category": "inference", "origin_date": "2023-09-29", "inventor_lab": "MIT" }, { "id": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "slug": "instruction-tuning-flan", "name": "Instruction Tuning (FLAN)", "type": "technique", "category": "training", "origin_date": "2021-09-03", "inventor_lab": "Google" }, { "id": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "slug": "chain-of-thought", "name": "Chain-of-Thought Prompting", "type": "technique", "category": "reasoning", "origin_date": "2022-01-28", "inventor_lab": "Google" }, { "id": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "slug": "zero-shot-cot", "name": "Zero-Shot Chain-of-Thought", "type": "technique", "category": "reasoning", "origin_date": "2022-05-24", "inventor_lab": "University of Tokyo" }, { "id": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "slug": "self-consistency", "name": "Self-Consistency", "type": "technique", "category": "reasoning", "origin_date": "2022-03-21", "inventor_lab": "Google" }, { "id": "9d60d0e7-5ac4-4f51-82f6-3299603249ce", "slug": "self-instruct", "name": "Self-Instruct", "type": "technique", "category": "training", "origin_date": "2022-12-20", "inventor_lab": "University of Washington" }, { "id": "30610518-977c-4c69-8c59-13a8fed66de0", "slug": "deep-rl-from-human-preferences", "name": "Deep RL from Human Preferences", "type": "technique", "category": "alignment", "origin_date": "2017-06-12", "inventor_lab": "OpenAI" }, { "id": "3e4551d3-eebc-4f6a-9707-7d44b7d64b07", "slug": "rlhf", "name": "Reinforcement Learning from Human Feedback (RLHF)", "type": "technique", "category": "alignment", "origin_date": "2022-03-04", "inventor_lab": "OpenAI" }, { "id": "d8765d34-96e5-4cad-ad58-bc2f741fe5cb", "slug": "dpo", "name": "Direct Preference Optimization (DPO)", "type": "technique", "category": "alignment", "origin_date": "2023-05-29", "inventor_lab": "Stanford" }, { "id": "53b6343e-ec96-4f12-8594-7b701ef835f9", "slug": "ipo", "name": "Identity Preference Optimization (IPO)", "type": "technique", "category": "alignment", "origin_date": "2023-10-18", "inventor_lab": "Google DeepMind" }, { "id": "4bbf88e8-4f58-4078-812b-dabbdb4f1897", "slug": "kto", "name": "KTO (Kahneman-Tversky Optimization)", "type": "technique", "category": "alignment", "origin_date": "2024-02-02", "inventor_lab": "Contextual AI" }, { "id": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "slug": "constitutional-ai", "name": "Constitutional AI", "type": "technique", "category": "alignment", "origin_date": "2022-12-15", "inventor_lab": "Anthropic" }, { "id": "2566df1a-ca40-414e-b3cf-ab88f0ed86c6", "slug": "rlaif", "name": "RLAIF (Reinforcement Learning from AI Feedback)", "type": "technique", "category": "alignment", "origin_date": "2023-09-01", "inventor_lab": "Google" }, { "id": "1ae70e23-c00c-4639-84f0-50059b00e87a", "slug": "qlora", "name": "QLoRA", "type": "technique", "category": "training", "origin_date": "2023-05-23", "inventor_lab": "University of Washington" }, { "id": "20bbaff7-e25b-412b-94bb-4e46f86f6b1c", "slug": "chinchilla-scaling", "name": "Chinchilla Scaling Laws", "type": "technique", "category": "training", "origin_date": "2022-03-29", "inventor_lab": "Google DeepMind" }, { "id": "9ef46e9d-87ec-4632-9567-a6610ef0293d", "slug": "react-agent", "name": "ReAct (Reason + Act)", "type": "technique", "category": "agents", "origin_date": "2022-10-06", "inventor_lab": "Princeton / Google" }, { "id": "6ee6b429-b6ec-48d3-bd6f-94d491089b72", "slug": "toolformer", "name": "Toolformer (Tool Use)", "type": "technique", "category": "agents", "origin_date": "2023-02-09", "inventor_lab": "Meta AI" }, { "id": "abaf391d-4f31-4e70-b8c7-bd9569eb2e09", "slug": "rag", "name": "Retrieval-Augmented Generation (RAG)", "type": "technique", "category": "retrieval", "origin_date": "2020-05-22", "inventor_lab": "Meta AI" }, { "id": "1e35a2b3-7e26-4b36-820d-59255b10ea1a", "slug": "dense-retrieval-dpr", "name": "Dense Passage Retrieval (DPR)", "type": "technique", "category": "retrieval", "origin_date": "2020-04-10", "inventor_lab": "Meta AI" }, { "id": "7727e97d-8663-4a11-a27a-06432105be7b", "slug": "tree-of-thoughts", "name": "Tree of Thoughts", "type": "technique", "category": "reasoning", "origin_date": "2023-05-17", "inventor_lab": "Princeton / Google DeepMind" }, { "id": "eaa22f4e-c263-4ba3-aaaf-d3eadf686a1f", "slug": "reflexion", "name": "Reflexion", "type": "technique", "category": "agents", "origin_date": "2023-03-20", "inventor_lab": "Northeastern / MIT" }, { "id": "9253adcc-5077-44c9-b892-03ecff6e6b8a", "slug": "process-reward-models", "name": "Process Reward Models", "type": "technique", "category": "reasoning", "origin_date": "2023-05-31", "inventor_lab": "OpenAI" }, { "id": "80779017-dadc-4b81-8eb2-2839e01df206", "slug": "test-time-compute-scaling", "name": "Test-Time Compute Scaling", "type": "technique", "category": "reasoning", "origin_date": "2024-08-06", "inventor_lab": "Google DeepMind" }, { "id": "442e2974-73ff-43e5-ac67-039e81c6fd2a", "slug": "rejection-sampling-ft", "name": "Rejection Sampling Fine-Tuning", "type": "technique", "category": "training", "origin_date": "2023-07-18", "inventor_lab": "Meta AI" }, { "id": "352dbc07-4786-4f18-97dc-f11d5262f777", "slug": "synthetic-data-orca", "name": "Synthetic Data Distillation (Orca)", "type": "technique", "category": "training", "origin_date": "2023-06-05", "inventor_lab": "Microsoft Research" }, { "id": "1b1a5cba-127c-461d-937c-6d9a6d51e0f2", "slug": "self-rewarding-lm", "name": "Self-Rewarding Language Models", "type": "technique", "category": "alignment", "origin_date": "2024-01-18", "inventor_lab": "Meta AI" }, { "id": "e14bfc5b-bd49-4464-9a7b-a129abe8c7a9", "slug": "sparse-autoencoders", "name": "Sparse Autoencoders for Interpretability", "type": "technique", "category": "interpretability", "origin_date": "2023-10-05", "inventor_lab": "Anthropic" }, { "id": "d416344e-bfed-4adb-9324-4ed9f4c93af9", "slug": "red-teaming-preference", "name": "Red-Teaming with Preference Models", "type": "technique", "category": "alignment", "origin_date": "2022-02-07", "inventor_lab": "Google DeepMind" }, { "id": "0c3b26ec-95ae-44b6-8bc7-6576dc8e49ac", "slug": "int8-llm-quantization", "name": "INT8 Weight Quantization for LLMs", "type": "technique", "category": "inference", "origin_date": "2022-08-15", "inventor_lab": "University of Washington" }, { "id": "9300fbea-66bf-47ee-97ec-4ed4e17c8bc5", "slug": "gptq", "name": "GPTQ Quantization", "type": "technique", "category": "inference", "origin_date": "2022-10-31", "inventor_lab": "ISTA" }, { "id": "9ff4bc90-e9fb-4a85-93f8-c298d2dc98f8", "slug": "awq-quantization", "name": "AWQ (Activation-Aware Weight Quantization)", "type": "technique", "category": "inference", "origin_date": "2023-06-01", "inventor_lab": "MIT" }, { "id": "9aa3f551-1ea1-47d5-9f2b-554499b1cc4a", "slug": "lora", "name": "LoRA (Low-Rank Adaptation)", "type": "technique", "category": "training", "origin_date": "2021-06-17", "inventor_lab": "Microsoft" }, { "id": "ca013160-ff76-42e4-9575-9ce8776653a2", "slug": "clip", "name": "CLIP (Contrastive Language-Image Pretraining)", "type": "technique", "category": "multimodal", "origin_date": "2021-02-26", "inventor_lab": "OpenAI" }, { "id": "4991d610-3d64-4b60-9bda-6e425f3e0a76", "slug": "latent-diffusion", "name": "Latent Diffusion", "type": "technique", "category": "multimodal", "origin_date": "2021-12-20", "inventor_lab": "LMU Munich / RunwayML" }, { "id": "80b7eef3-60a8-4060-becc-289f6f22c4a2", "slug": "vit", "name": "Vision Transformer (ViT)", "type": "technique", "category": "multimodal", "origin_date": "2020-10-22", "inventor_lab": "Google" }, { "id": "bcc4e0b4-8b7a-4bb0-8589-acf0131b9539", "slug": "flamingo-vlm", "name": "Flamingo (Cross-Attention VLMs)", "type": "technique", "category": "multimodal", "origin_date": "2022-04-29", "inventor_lab": "Google DeepMind" }, { "id": "a64d7077-9ad0-42c0-93f7-1cdf2c13899f", "slug": "llava-vit", "name": "LLaVA (Visual Instruction Tuning)", "type": "technique", "category": "multimodal", "origin_date": "2023-04-17", "inventor_lab": "University of Wisconsin" }, { "id": "0c27a9a7-3728-468d-80ee-6e19b4fdd1f0", "slug": "whisper-asr", "name": "Whisper (Robust Speech Recognition)", "type": "technique", "category": "multimodal", "origin_date": "2022-12-06", "inventor_lab": "OpenAI" }, { "id": "046912ba-66f3-4f53-ab94-7106859f148a", "slug": "kimi-k2-5", "name": "Kimi K2.5", "type": "product", "company": null, "first_seen": "2026-03-04T20:03:25.096865+00:00" }, { "id": "17338571-ab3f-4843-bc61-2988824bc78a", "slug": "glm-5-1", "name": "GLM-5.1", "type": "product", "company": null, "first_seen": "2026-03-21T16:49:37.113993+00:00" }, { "id": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "slug": "deepseek-r1", "name": "DeepSeek-R1", "type": "product", "company": null, "first_seen": "2026-03-17T22:04:13.040500+00:00" }, { "id": "224b5abf-7eaa-43b7-9c9d-d56b6086f0ea", "slug": "mistral-small-4", "name": "Mistral Small 4", "type": "product", "company": null, "first_seen": "2026-03-16T23:45:13.789534+00:00" }, { "id": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "slug": "gemini-3-flash", "name": "Gemini 3 Flash", "type": "product", "company": null, "first_seen": "2026-02-27T19:28:23.064338+00:00" }, { "id": "51508449-e877-4518-986a-c4d9729ab379", "slug": "deepseek-v3", "name": "DeepSeek-V3", "type": "product", "company": null, "first_seen": "2026-03-11T05:34:48.533814+00:00" }, { "id": "5512bb31-f60b-4274-8e57-a19ca3f98daa", "slug": "llama-4-scout", "name": "Llama 4 Scout", "type": "product", "company": null, "first_seen": "2025-04-05T00:00:00+00:00" }, { "id": "59ee160b-5560-4067-b597-70eaaec27c4c", "slug": "gpt-5-2-pro", "name": "GPT-5.2 Pro", "type": "product", "company": null, "first_seen": "2026-02-17T16:12:10.857388+00:00" }, { "id": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "slug": "gpt-4o", "name": "GPT-4o", "type": "product", "company": null, "first_seen": "2026-02-16T11:01:17.872655+00:00" }, { "id": "5de2907a-ba1b-499b-bb93-3a0eca1f5042", "slug": "nemotron-3-super", "name": "Nemotron 3 Super", "type": "product", "company": null, "first_seen": "2026-03-11T17:49:02.200699+00:00" }, { "id": "66751db7-9306-4304-9dec-16d0f686a1c8", "slug": "claude-sonnet-4-6", "name": "Claude Sonnet 4.6", "type": "product", "company": null, "first_seen": "2026-02-25T00:49:56.536486+00:00" }, { "id": "6708b6a1-0856-4bb6-a654-c0dc0c964893", "slug": "gpt-oss-120b", "name": "GPT-OSS-120B", "type": "product", "company": null, "first_seen": "2026-03-02T14:32:55.710837+00:00" }, { "id": "71461b72-6b33-412e-be5c-5323232e1c5c", "slug": "gpt-5-4-cyber", "name": "GPT-5.4-Cyber", "type": "product", "company": null, "first_seen": "2026-04-16T18:11:33.003755+00:00" }, { "id": "74c6bb51-637d-43b7-aad6-a333754ffc06", "slug": "claude-3-5-opus", "name": "Claude 3.5 Opus", "type": "product", "company": null, "first_seen": "2026-03-18T15:05:59.021685+00:00" }, { "id": "77f12f13-f612-4b03-ad74-5825c3696778", "slug": "gpt-rosalind", "name": "GPT-Rosalind", "type": "product", "company": null, "first_seen": "2026-04-16T20:49:04.321574+00:00" }, { "id": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "slug": "gpt-5-3", "name": "GPT-5.3", "type": "product", "company": null, "first_seen": "2026-02-26T11:15:47.694912+00:00" }, { "id": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "slug": "claude-3", "name": "Claude 3", "type": "product", "company": null, "first_seen": "2026-02-18T12:35:38.233947+00:00" }, { "id": "963f7812-c831-44e3-9e70-2e9cb77a7dbe", "slug": "llama-4-maverick", "name": "Llama 4 Maverick", "type": "product", "company": null, "first_seen": "2025-04-05T00:00:00+00:00" }, { "id": "a66e5f41-316f-42c1-a062-4bbe9b1f22ea", "slug": "grok-4-20", "name": "Grok 4.20", "type": "product", "company": null, "first_seen": "2026-03-02T20:06:04.692632+00:00" }, { "id": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "slug": "gemini-3-pro", "name": "Gemini 3 Pro", "type": "product", "company": null, "first_seen": "2026-02-19T10:18:10.633061+00:00" }, { "id": "b6a1a236-7380-4999-8748-20e6bb8bac76", "slug": "claude-haiku-4-5", "name": "Claude Haiku 4.5", "type": "product", "company": null, "first_seen": "2026-02-25T00:49:56.546879+00:00" }, { "id": "bcacead1-07af-456f-ad4e-57b90246eae8", "slug": "gemini-3-1-flash-lite", "name": "Gemini 3.1 Flash-Lite", "type": "product", "company": null, "first_seen": "2026-03-05T12:21:36.323569+00:00" }, { "id": "bcdb6718-22ce-4309-9fd4-d7f5840d2996", "slug": "claude-opus-4-6", "name": "Claude Opus 4.6", "type": "product", "company": null, "first_seen": "2026-02-16T11:01:17.857653+00:00" }, { "id": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "slug": "claude-4-5", "name": "Claude 4.5", "type": "product", "company": null, "first_seen": "2026-02-26T11:15:47.716542+00:00" }, { "id": "cae56539-eb7a-4fb1-a1d4-e57f88239b6e", "slug": "claude-opus-4-7", "name": "Claude Opus 4.7", "type": "product", "company": null, "first_seen": "2026-04-16T22:20:47.591836+00:00" }, { "id": "d54f97b3-9486-40e6-b80c-7e3aba39c634", "slug": "claude-3-5-sonnet", "name": "Claude 3.5 Sonnet", "type": "product", "company": null, "first_seen": "2026-02-23T04:36:30.024632+00:00" }, { "id": "e3a50221-7840-4da7-b805-f65352c08329", "slug": "kimi-k2-6", "name": "Kimi K2.6", "type": "product", "company": null, "first_seen": "2026-04-20T00:00:00+00:00" }, { "id": "e62ec052-492b-4177-b6e0-9b643812a104", "slug": "qwen-3-6", "name": "Qwen 3.6", "type": "product", "company": null, "first_seen": "2026-03-31T01:57:56.330620+00:00" }, { "id": "f161c02e-6361-4f16-ad71-619fadbe7c4a", "slug": "gemini-3-1", "name": "Gemini 3.1", "type": "product", "company": null, "first_seen": "2026-02-20T01:39:16.897479+00:00" }, { "id": "f398a219-1682-446d-a9f0-5acc72eb6f28", "slug": "gpt-5", "name": "GPT-5", "type": "product", "company": null, "first_seen": "2026-02-16T11:01:17.899226+00:00" } ], "edges": [ { "source": "cae56539-eb7a-4fb1-a1d4-e57f88239b6e", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "cae56539-eb7a-4fb1-a1d4-e57f88239b6e", "target": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "type": "deploys", "confidence": 0.9 }, { "source": "cae56539-eb7a-4fb1-a1d4-e57f88239b6e", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "cae56539-eb7a-4fb1-a1d4-e57f88239b6e", "target": "80779017-dadc-4b81-8eb2-2839e01df206", "type": "deploys", "confidence": 0.9 }, { "source": "cae56539-eb7a-4fb1-a1d4-e57f88239b6e", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "bcdb6718-22ce-4309-9fd4-d7f5840d2996", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "bcdb6718-22ce-4309-9fd4-d7f5840d2996", "target": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "type": "deploys", "confidence": 0.9 }, { "source": "bcdb6718-22ce-4309-9fd4-d7f5840d2996", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "bcdb6718-22ce-4309-9fd4-d7f5840d2996", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "bcdb6718-22ce-4309-9fd4-d7f5840d2996", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "type": "deploys", "confidence": 0.9 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "abaf391d-4f31-4e70-b8c7-bd9569eb2e09", "type": "deploys", "confidence": 0.9 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.6 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "c79e98ee-79f3-47ab-9976-bef7632d6d5d", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "66751db7-9306-4304-9dec-16d0f686a1c8", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.6 }, { "source": "66751db7-9306-4304-9dec-16d0f686a1c8", "target": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "type": "deploys", "confidence": 0.9 }, { "source": "66751db7-9306-4304-9dec-16d0f686a1c8", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.6 }, { "source": "66751db7-9306-4304-9dec-16d0f686a1c8", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "b6a1a236-7380-4999-8748-20e6bb8bac76", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "b6a1a236-7380-4999-8748-20e6bb8bac76", "target": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "type": "deploys", "confidence": 0.9 }, { "source": "b6a1a236-7380-4999-8748-20e6bb8bac76", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "b6a1a236-7380-4999-8748-20e6bb8bac76", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "b6a1a236-7380-4999-8748-20e6bb8bac76", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "d54f97b3-9486-40e6-b80c-7e3aba39c634", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "d54f97b3-9486-40e6-b80c-7e3aba39c634", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.6 }, { "source": "d54f97b3-9486-40e6-b80c-7e3aba39c634", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "d54f97b3-9486-40e6-b80c-7e3aba39c634", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "d54f97b3-9486-40e6-b80c-7e3aba39c634", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "d54f97b3-9486-40e6-b80c-7e3aba39c634", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "74c6bb51-637d-43b7-aad6-a333754ffc06", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "74c6bb51-637d-43b7-aad6-a333754ffc06", "target": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "type": "deploys", "confidence": 0.9 }, { "source": "74c6bb51-637d-43b7-aad6-a333754ffc06", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "74c6bb51-637d-43b7-aad6-a333754ffc06", "target": "9ef46e9d-87ec-4632-9567-a6610ef0293d", "type": "deploys", "confidence": 0.6 }, { "source": "74c6bb51-637d-43b7-aad6-a333754ffc06", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.6 }, { "source": "74c6bb51-637d-43b7-aad6-a333754ffc06", "target": "80779017-dadc-4b81-8eb2-2839e01df206", "type": "deploys", "confidence": 0.6 }, { "source": "74c6bb51-637d-43b7-aad6-a333754ffc06", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "e8ec8544-9749-4bf6-ad4e-c8ed463c3fe7", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "7f919848-a2a4-48ca-8e1b-8a955fa5ddcc", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "9ef46e9d-87ec-4632-9567-a6610ef0293d", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "3e4551d3-eebc-4f6a-9707-7d44b7d64b07", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.6 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "7a6ae0c1-370a-4cc7-8536-24b538de1320", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.6 }, { "source": "59ee160b-5560-4067-b597-70eaaec27c4c", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "59ee160b-5560-4067-b597-70eaaec27c4c", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "59ee160b-5560-4067-b597-70eaaec27c4c", "target": "3e4551d3-eebc-4f6a-9707-7d44b7d64b07", "type": "deploys", "confidence": 0.9 }, { "source": "59ee160b-5560-4067-b597-70eaaec27c4c", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.6 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.9 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "71461b72-6b33-412e-be5c-5323232e1c5c", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "77f12f13-f612-4b03-ad74-5825c3696778", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "77f12f13-f612-4b03-ad74-5825c3696778", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.9 }, { "source": "77f12f13-f612-4b03-ad74-5825c3696778", "target": "80779017-dadc-4b81-8eb2-2839e01df206", "type": "deploys", "confidence": 0.9 }, { "source": "77f12f13-f612-4b03-ad74-5825c3696778", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.6 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.6 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.6 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "f398a219-1682-446d-a9f0-5acc72eb6f28", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.6 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.6 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.6 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "f1d61976-639f-4df6-b2d2-7bc16dbaa86b", "type": "deploys", "confidence": 0.3 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "5ba56ca7-88e2-47d1-9c1e-2037b790ff54", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.9 }, { "source": "6708b6a1-0856-4bb6-a654-c0dc0c964893", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.6 }, { "source": "6708b6a1-0856-4bb6-a654-c0dc0c964893", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "6708b6a1-0856-4bb6-a654-c0dc0c964893", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "80b7eef3-60a8-4060-becc-289f6f22c4a2", "type": "deploys", "confidence": 0.9 }, { "source": "b38d2f39-311c-44c8-a9c1-4ac605d14ce0", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.9 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.9 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.9 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.6 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "506d5ce2-95b9-4bc4-9321-fa2d35549f3f", "target": "80b7eef3-60a8-4060-becc-289f6f22c4a2", "type": "deploys", "confidence": 0.9 }, { "source": "f161c02e-6361-4f16-ad71-619fadbe7c4a", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "f161c02e-6361-4f16-ad71-619fadbe7c4a", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.6 }, { "source": "f161c02e-6361-4f16-ad71-619fadbe7c4a", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.6 }, { "source": "f161c02e-6361-4f16-ad71-619fadbe7c4a", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "f161c02e-6361-4f16-ad71-619fadbe7c4a", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "f161c02e-6361-4f16-ad71-619fadbe7c4a", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.6 }, { "source": "bcacead1-07af-456f-ad4e-57b90246eae8", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "bcacead1-07af-456f-ad4e-57b90246eae8", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "bcacead1-07af-456f-ad4e-57b90246eae8", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "bcacead1-07af-456f-ad4e-57b90246eae8", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "bcacead1-07af-456f-ad4e-57b90246eae8", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "bcacead1-07af-456f-ad4e-57b90246eae8", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "a64d7077-9ad0-42c0-93f7-1cdf2c13899f", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "80b7eef3-60a8-4060-becc-289f6f22c4a2", "type": "deploys", "confidence": 0.6 }, { "source": "046912ba-66f3-4f53-ab94-7106859f148a", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.6 }, { "source": "e3a50221-7840-4da7-b805-f65352c08329", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "e3a50221-7840-4da7-b805-f65352c08329", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "e3a50221-7840-4da7-b805-f65352c08329", "target": "9ef46e9d-87ec-4632-9567-a6610ef0293d", "type": "deploys", "confidence": 0.9 }, { "source": "e3a50221-7840-4da7-b805-f65352c08329", "target": "80779017-dadc-4b81-8eb2-2839e01df206", "type": "deploys", "confidence": 0.9 }, { "source": "5512bb31-f60b-4274-8e57-a19ca3f98daa", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "5512bb31-f60b-4274-8e57-a19ca3f98daa", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "5512bb31-f60b-4274-8e57-a19ca3f98daa", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "5512bb31-f60b-4274-8e57-a19ca3f98daa", "target": "a64d7077-9ad0-42c0-93f7-1cdf2c13899f", "type": "deploys", "confidence": 0.6 }, { "source": "5512bb31-f60b-4274-8e57-a19ca3f98daa", "target": "ca013160-ff76-42e4-9575-9ce8776653a2", "type": "deploys", "confidence": 0.6 }, { "source": "963f7812-c831-44e3-9e70-2e9cb77a7dbe", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "963f7812-c831-44e3-9e70-2e9cb77a7dbe", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "963f7812-c831-44e3-9e70-2e9cb77a7dbe", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "963f7812-c831-44e3-9e70-2e9cb77a7dbe", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.6 }, { "source": "224b5abf-7eaa-43b7-9c9d-d56b6086f0ea", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "224b5abf-7eaa-43b7-9c9d-d56b6086f0ea", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "224b5abf-7eaa-43b7-9c9d-d56b6086f0ea", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.9 }, { "source": "224b5abf-7eaa-43b7-9c9d-d56b6086f0ea", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.9 }, { "source": "224b5abf-7eaa-43b7-9c9d-d56b6086f0ea", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.9 }, { "source": "224b5abf-7eaa-43b7-9c9d-d56b6086f0ea", "target": "91fa092e-f841-4016-93a8-7f2716090cc0", "type": "deploys", "confidence": 0.9 }, { "source": "51508449-e877-4518-986a-c4d9729ab379", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "51508449-e877-4518-986a-c4d9729ab379", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "51508449-e877-4518-986a-c4d9729ab379", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.9 }, { "source": "51508449-e877-4518-986a-c4d9729ab379", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.9 }, { "source": "51508449-e877-4518-986a-c4d9729ab379", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.9 }, { "source": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.9 }, { "source": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "target": "80779017-dadc-4b81-8eb2-2839e01df206", "type": "deploys", "confidence": 0.9 }, { "source": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "target": "3e4551d3-eebc-4f6a-9707-7d44b7d64b07", "type": "deploys", "confidence": 0.9 }, { "source": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "target": "9253adcc-5077-44c9-b892-03ecff6e6b8a", "type": "deploys", "confidence": 0.9 }, { "source": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "1c6bc2fb-4b04-4706-ba35-da10dda415f9", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "a66e5f41-316f-42c1-a062-4bbe9b1f22ea", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.6 }, { "source": "a66e5f41-316f-42c1-a062-4bbe9b1f22ea", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.6 }, { "source": "a66e5f41-316f-42c1-a062-4bbe9b1f22ea", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "a64d7077-9ad0-42c0-93f7-1cdf2c13899f", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "9aa3f551-1ea1-47d5-9f2b-554499b1cc4a", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.6 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "80b7eef3-60a8-4060-becc-289f6f22c4a2", "type": "deploys", "confidence": 0.9 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.6 }, { "source": "e62ec052-492b-4177-b6e0-9b643812a104", "target": "2cc8949b-f306-4290-a34b-b20f1d26f19f", "type": "deploys", "confidence": 0.6 }, { "source": "5de2907a-ba1b-499b-bb93-3a0eca1f5042", "target": "30e88a7b-e2b0-4900-905a-9a82f729f4fa", "type": "deploys", "confidence": 0.9 }, { "source": "5de2907a-ba1b-499b-bb93-3a0eca1f5042", "target": "644f3859-e61a-4bd2-bb25-6a43f40a8b4a", "type": "deploys", "confidence": 0.9 }, { "source": "5de2907a-ba1b-499b-bb93-3a0eca1f5042", "target": "f41e091c-2cdb-4589-9734-ced31b18b372", "type": "deploys", "confidence": 0.9 }, { "source": "17338571-ab3f-4843-bc61-2988824bc78a", "target": "53b1c32a-4b47-4077-8c67-81060cbb6420", "type": "deploys", "confidence": 0.9 }, { "source": "17338571-ab3f-4843-bc61-2988824bc78a", "target": "b79a0be1-b1e0-40f1-8574-d010e1195da7", "type": "deploys", "confidence": 0.9 }, { "source": "17338571-ab3f-4843-bc61-2988824bc78a", "target": "8608614e-381c-464e-a424-8ce991c97ef4", "type": "deploys", "confidence": 0.9 }, { "source": "17338571-ab3f-4843-bc61-2988824bc78a", "target": "f10822f0-23df-4bde-86fb-d4180e9fe2cd", "type": "deploys", "confidence": 0.9 }, { "source": "17338571-ab3f-4843-bc61-2988824bc78a", "target": "ebfb4ee8-85e0-47ee-b402-9d6fa44437cd", "type": "deploys", "confidence": 0.6 }, { "source": "17338571-ab3f-4843-bc61-2988824bc78a", "target": "ca8cdb04-64fc-4fce-8d55-c96eaaaa7184", "type": "deploys", "confidence": 0.6 }, { "source": "17338571-ab3f-4843-bc61-2988824bc78a", "target": "367a29b2-0553-4c7e-9d32-88a1bfc26da2", "type": "deploys", "confidence": 0.6 } ] } }