会员服务
CARE-PO: Controllable Tool-Call Efficiency for LLM Agents
Arxiv
0+阅读 · 11月9日
Speculative Self-Distillation enables Efficient Knowledge Internalization
Arxiv
0+阅读 · 11月9日
HiFloat4 Format for End-To-End Reinforcement Learning Post-Training of Large Language Models
Arxiv
0+阅读 · 11月9日
Watching Itself Watch: Self-Auditing Visual Reliance for Video Reasoning
Arxiv
0+阅读 · 11月9日
Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
Arxiv
0+阅读 · 11月9日
Adaptive Scheduling Pipeline for Multi-Instance Asynchronous Reinforcement Learning
Arxiv
0+阅读 · 11月9日
How Many Candidates Does a One-Step World Model Need? A Random-Codebook Account of Nearest-of-m Training on Frozen DINO Tokens
Arxiv
0+阅读 · 11月9日
PhysOPSD: Code as Privilege for Physically Plausible Video Generation
Arxiv
0+阅读 · 11月9日
Prompt Driven Exploration: Language as an Exploration Space for VLA Reinforcement Learning
Arxiv
0+阅读 · 11月8日
BunshinRL: Splitting Live Agent Workspaces for Memory-Efficient Agentic RL
Arxiv
0+阅读 · 11月7日
Dynamics-Informed Action Learning for Goal-Conditioned Latent Planning
Arxiv
0+阅读 · 11月7日
MoE-DAgger: Transition-Aware Router-Entropy Gating for Selective Human Correction in Phase-Structured Manipulation
Arxiv
0+阅读 · 11月7日
BRAT: Bounded Residual-Aligned Training for Stable Multi-Pass LLM Post-Training
Arxiv
0+阅读 · 11月7日
Detent: Batch-Invariant LLM Serving and RL Rollouts at Near-Stock Speed
Arxiv
0+阅读 · 11月7日
RELAY: Selective Self-Distillation with Adaptive Hand-off for Rubric-Based Reinforcement Learning
Arxiv
0+阅读 · 11月7日
参考链接
微信扫码咨询专知VIP会员
Top