llm
30 articles about llm in AI news
HAVEN Benchmark Exposes MLLM Gap Between Fluency and Video Understanding
HAVEN benchmark tests MLLMs on hierarchical video understanding across frame, shot, and video levels. Results show top models lack grounded multimodal reasoning despite fluent text generation.
Memory as a Model: Augmenting LLMs with Trained Memory
Paper augments LLMs with trained memory for long-term recall. Model-agnostic approach stores external knowledge without retraining.
OpenAI Readies General-Purpose LLM With Test-Time Compute Scaling
OpenAI is releasing a general-purpose LLM that improves with test-time compute, per an internal message. The model shows math gains without specialized training.
Apple Paper Argues LLMs Show 'Illusion of Thinking'
Apple paper argues LLMs show no genuine reasoning, only pattern matching. The critique targets vendor claims but lacks new empirical evidence.
train-llm-from-scratch: 1B-Parameter LLM on a Single GPU
train-llm-from-scratch trains billion-parameter LLMs on a single GPU, cutting costs from $10M+ to consumer hardware.
Persuasion Techniques Boost LLM Compliance from 35% to 51% in PNAS Study
PNAS study finds persuasion techniques boost LLM compliance from 35% to 51%, with newer models resisting more.
MLLM Raters Show Central Tendency Bias in Clinical Scoring
Study finds GPT-5 and other MLLMs show central tendency bias in clinical scoring, compressing predictions toward scale midpoint despite prompt modifications.
LLM-EDT: Dual-Phase Training Boosts Cross-Domain Rec by 12.4%
LLM-EDT improves cross-domain sequential recommendation by up to 12.4% using dual-phase training and LLM-based item generation.
Cascaded LLMs Lift E-Commerce Cart Adds 2.7% in Online Test
A cascaded LLM framework for e-commerce storefront generation lifted cart adds by +2.7% in online tests, using teacher-student fine-tuning to approach closed-weight LLM quality at production latency.
vLLM Optimizations Cut Voice AI Latency by 40% on 6-GPU Cluster
vLLM optimizations on a 6-GPU cluster reduced voice AI latency by 40% for a Qwen-based system, enabling 500 concurrent sessions per node without hardware upgrades.
SDAR: Self-Distilled RL Stabilizes Multi-Turn LLM Agents, +9.4% on ALFWorld
SDAR gates self-distillation within GRPO to stabilize multi-turn LLM agent training, yielding +9.4% on ALFWorld and gains on WebShop and Search-QA across Qwen2.5 and Qwen3 models.
Collider-Bench Tests LLM Agents on LHC Analysis Reproduction
Collider-Bench tests LLM agents on reproducing LHC analyses from papers. No agent beats physicist-in-the-loop, highlighting gaps in scientific reasoning.
VAB Benchmark: Top MLLMs Judge Beauty Correctly Only 26.5% of Time
Frontier MLLMs achieve only 26.5% accuracy on VAB, far below human 68.9%. Fine-tuning bridges the gap.
LLM Pipelines Beat Regex at Invoice Extraction at Scale
LLM pipelines outperform regex for structured extraction from unstructured documents, handling 20+ invoice formats without per-format rule maintenance.
Multi-Agent LLM Systems Fail to Outperform Single Models, Study Finds
New paper finds multi-agent LLM systems underperform single models by 2.3% on reasoning benchmarks, challenging a core assumption in AI engineering.
Pruning LLMs for Edge Triples Bias, Perplexity Hides Damage
Pruning LLMs for edge deployment amplifies bias up to 83.7% while perplexity barely changes, revealing a paradox that undermines standard evaluation practices.
MM-LLM Framework Boosts Recommendation AUC 0.35%, Online Metrics 0.02%
arXiv paper proposes LLaMA2-based MM-LLM framework for recommendation, achieving 0.35% AUC gain and 0.02% online lift at scale.
OSA Injects Ordinal Semantics into LLM Recommenders, Beats CF Baselines
OSA injects ordinal semantics into LLM-based recommenders using token embeddings as anchors, outperforming prior CF-LLM methods on pairwise preference evaluation.
SalesSim: LLMs Score Below 79% on Retail Persona Alignment, RL Boosts 13.8%
SalesSim benchmarks MLLMs as retail customers; top models score below 79% on persona alignment. UserGRPO RL boosts alignment by 13.8%.
RRCM Uses GRPO to Decide When to Retrieve for LLM Recommendation
RRCM uses GRPO to learn when to retrieve evidence for LLM recommendation, outperforming fixed-context baselines.
Two-Tower vs Vector DB + LLM: Which Wins for RecSys at Scale?
Two-tower models offer sub-10ms latency for cold-start; vector DB + LLM provides richer semantics. Hybrid architectures reduce churn by 15-20%.
Claude Code's HTML Output Beats Markdown for LLM-Readable Docs
Claude Code generates HTML docs that LLMs parse more accurately than Markdown, per Thariq's analysis. Trade-off: harder for humans to edit.
LLMs Fail at Implicit Travel Constraints, New Benchmark Shows
LLMs fail at implicit travel constraints, a new arXiv paper decomposes planning into 5 atomic skills, finding structural biases and ineffective self-correction.
Unsloth × NVIDIA Cut LLM Fine-Tuning ~25% — Three Glue-Code Wins on Blackwell
Daniel & Michael Han at Unsloth, in collaboration with NVIDIA, published a joint guide quantifying three glue-code optimizations that combine for ~25% faster LLM training on B200 Blackwell hardware. The wins target overhead around the main kernels — caching packed-sequence metadata, double-buffered gradient checkpoint reloads, and a cheaper GPT-OSS MoE router using argsort + bincount. All three are merged via public PRs.
ARMOR 2025: Military Safety Benchmark Exposes LLM Gaps Across 21 Models
ARMOR 2025 benchmark tests 21 LLMs against military legal doctrines, revealing critical safety gaps that civilian benchmarks miss.
Microsoft: LLMs Corrupt 25% of Docs in Long Edits
Microsoft paper shows LLMs corrupt ~25% of documents across 52 domains during 20-edit sessions, with failures compounding silently.
LLMs Shrink Neural Activity When Confused, New Paper Shows
LLMs compress neural activity when confused, measurable as a sparsity signal. Paper 2603.03415 proposes using this for adaptive prompting.
K-CARE: A New Framework Grounds LLMs in External Knowledge to Fix
K-CARE combines Symmetrical Contextual Anchoring (behavior data) and Analogical Prototype Reasoning (expert examples) to resolve e-commerce search relevance issues that pure LLM reasoning can't fix. Proven in offline and online A/B tests on a leading platform.
Vibe Training: SLM Replaces LLM-as-a-Judge, 8x Faster, 50% Fewer Errors
Plurai introduces 'vibe training,' using adversarial agent swarms to distill a small language model (SLM) for evaluating and guarding production AI agents. The SLM outperforms standard LLM-as-a-judge setups with ~8x faster inference and ~50% fewer evaluation errors.
KARL: RL Framework Cuts LLM Hallucinations Without Accuracy Loss
KARL introduces a reinforcement learning framework that dynamically estimates an LLM's knowledge boundary to reward abstention only when appropriate, achieving a superior accuracy-hallucination trade-off on multiple benchmarks without sacrificing correctness.