A new arXiv study (2608.20362v1) finds that exact-match verifiers in reinforcement learning with verifiable rewards (RLVR) introduce language-dependent false-negative reward noise in multilingual sett…
Researchers propose DUET, a token-selective on-policy distillation method that uses two identical-weight teachers differing only in prohibition visibility to improve LLM compliance with runtime-inject…
Researchers introduced Attention-Aware Transform Coding (AATC), a KV cache compression method that reduces memory use by approximately 5.8x while maintaining near-lossless accuracy on models like Llam…
Researchers propose Thought-Aware Attention Matching (TAM), a KV cache compaction method that segments chain-of-thought reasoning into blocks, allocates compression budgets adaptively, and protects pi…
At YCML, YC's first machine learning research showcase at Startup School 2026, eight researchers presented papers on topics including model reasoning, formal mathematics, video agents, and robotics. N…
LOCKS, a new method from arXiv, enables efficient long-context decoding by giving each page of the key-value cache its own compact spectral summary, reconstructing within-page logits, and attending on…
Kimi K3 is 30x cheaper than GPT-5 for output tokens while leading the LMArena leaderboard with 1,289 ELO, according to a 2026 comparison. For a customer support bot handling 10M input and 5M output to…
Researchers propose LISA (Linear-Indexed Sparse Attention), a plug-and-play attention module that reduces inference complexity from O(n^2) to O(nM) for long-context reasoning models. Experiments on De…
Researchers propose BIRD (Bootstrapped Iterative Self-Reasoning Distillation), a two-stage method that improves compressed reasoning by first generating concise, correct traces from a base model under…
SPARK, a new approach to diagnosing hidden-state failures in AI reasoning, boosted accuracy on the MATH-500 benchmark for Qwen3-4B from 82.0% to 84.6% and for Qwen3-8B from 82.4% to 85.6%, according t…
A new study evaluating 67 frontier AI models from 21 providers reveals that enterprises underestimate AI model failure rates by 2.25 to 2.5 times due to a mathematical limit called the "co-failure cei…
Researchers propose R²PO, a method that decouples rollout and inference policies for LLM reasoning, achieving accuracy gains of 3.4% on MATH-500 and 1.3% on APPS. The approach attaches a lightweight R…
Researchers introduced LearnStop, a cost-aware early-exit method for reasoning language models that learns when to stop computation based on online features like answer confidence and entropy. Across …
A developer recounts pushing an agent into production that failed after a CRM dropdown change, highlighting the inadequacy of model-level evaluation for agent systems. The article argues that agents m…
Researchers introduced JetSpec, a speculative decoding method that trains a causal parallel draft head on a frozen target model to draft entire speculative trees in one pass while preserving autoregre…
JetSpec, a new speculative decoding method, trains a causal parallel draft head over fused hidden states from a frozen target model, enabling lossless verification of candidate trees in one forward pa…
Researchers introduced GRACE, a theoretical framework that determines the optimal verification granularity for test-time scaling in large language models based on problem difficulty, verifier accuracy…
Researchers from Hao AI Lab introduced JetFlow, a speculative decoding framework that breaks the scaling ceiling of autoregressive LLMs by combining one-forward drafting efficiency with branch-wise ca…
Researchers at UC Riverside introduced the first robustness study for proof autoformalization in Lean 4, testing seven LLM-based models on perturbed informal proofs. All models showed sensitivity to g…
Researchers introduced Sequential Bayesian Belief Tracking (SBBT), a framework that estimates the likelihood of a correct final answer from partial reasoning traces by calibrating observation likeliho…