Category
7,776 papers
Temperature-Grouped Reinforcement Learning (TGRL), which turns temperature-induced diversity into an explicit training signal, and reaches equivalent accuracy up to 36% faster than strong RLVR baselines without expanding the rollout budget.
SMAT (Simple MAT) is introduced, which jointly optimizes expert loss and expected loss at simulated merged parameters generated by sampling scaling coefficients, masks, and additive noise, and improves the mean score across five merging methods.
CalibHyper is proposed, a chance-corrected relational hypergraph method based on the joint label distribution that achieves ROC-AUC competitive with the strongest reported results.
Training-free KV-cache compression methods such as SnapKV, H2O, and PyramidKV evict tokens at the end of prefill, aiming to preserve the attention mass that future queries are expected to use -optimizing what to keep. We show that this objective fails in two distinct ways. (1) Compensation: restoring the evicted attent...
Reach audiences
Advertise in front of researchers, engineers, and readers.
On-Policy Attention Self-Distillation (OPASD), which complements token-level supervision with solution-conditioned attention distillation, shows that solution-conditioned attention provides a complementary supervision signal that makes on-policy self-distillation more accurate, stable, and compute-efficient.
This work studies language-model synthesis of Clifford circuits, which prepare the stabilizer states used in quantum error correction and admit exact classical verification, in a target-conditioned framework.
Group-relative reinforcement learning (RL) relies on reward variation among sampled responses to estimate informative relative advantages. As language models become increasingly capable, existing training data can become reward-saturated: all sampled responses to the same problem might receive equally high rewards, whe...
Tory of Scene (ToS), a training-free reasoning schema in which each agent reads its public role, the only difference between the agents, and the task context they all observe, is proposed, which outperforms all six baselines on every benchmark, and each baseline falls far behind it in at least one setting.
LLM post-training combines supervised fine-tuning (SFT), a mode-covering forward-KL objective, with reinforcement learning (RL), a mode-seeking reverse-KL objective, with reinforcement learning (RL), a mode-seeking reverse-KL objective to exploit anisotropy, a division of labor that post-training can exploit.
The memory usage and decoding latency of LLM inference grow rapidly with context length. To reduce these costs, key-value (KV) cache compression methods selectively retain cached states based on token importance or differences in attention patterns across heads. However, we discover that retrieval capability varies sub...
Large language models can generate several records within one autoregressive completion, making earlier answers available as context for later records. This paper shows that such shared-completion batching can distort relationships among variables in the resulting synthetic data, using controlled tests on synthetic sur...
HM-Router is introduced, a routing method that jointly selects a model and harness for each query, with an interaction term inspired by canonical polyadic tensor decomposition to capture how their compatibility varies with the query.
A new computational approach reveals how cells communicate to assemble into tissues. This could accelerate stem cell engineering, improve disease models, and inform therapies.
Zero-knowledge proofs can verify computing results without rerunning workloads or exposing sensitive data. Magnetar helps bring them into the mainstream with support for standard programming languages, major speed gains, and quantum-resistant security. The post Proof-based general-purpose computing with Magnetar appeared first on Microsoft Research.
A new language-processing tool could help identify the highest-risk individuals from natural language, enabling swifter interventions.
A Blog post by Liquid AI on Hugging Face
Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking are our most advanced live dialogue models yet, built for natural conversation.
A new method, called CW-Net, translates the reasoning process of an autonomous vehicle’s AI system into understandable concepts that explain its behavior.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
