BrowseComp-Plus

Headlines · 2

1. Oct 23
   CLM lets language models autonomously manage their context

   InfoQ 中文 AI&大模型 reports that researchers from Meta, MIT, and the University of Washington proposed Context Language Model (CLM), which treats context as an editable file and can rewrite messages, preserve important facts, remove irrelevant information, and maintain progress notes. The researchers evaluated zero-shot, in-context, and reinforcement-learning approaches; reported results included an 11.4% accuracy gain and 21.5% fewer FLOPs on BrowseComp-Plus, while the reinforcement-learning variant raised Qwen3.5-9B accuracy from 28.8% to 42.5% with 12% fewer FLOPs. The report also notes risks including lost information, persistent prompt injection or self-generated instructions, cache invalidation, and incorrect context-management decisions.

   InfoQ 中文 AI&大模型 · 🔥 0
2. Oct 19
   Self-Credit paper adds dense credit assignment to long-horizon agentic RL

   A new arXiv paper (2610.28388, submitted 20 Oct 2026) by Fangcong Yin, Ming Yin and colleagues proposes Self-Credit, an end-to-end agentic reinforcement learning method that turns sparse outcome rewards into dense supervision signals via self-generated credit assignment and intra-group rescaling. The method builds on group-based RL with three parts: the policy model writes summaries citing earlier turns that held critical evidence, intra-group rescaling compares cited actions across successful and failed rollouts of the same prompt, and advantage reweighting applies the rescaled weights to the cited turns. The authors report that Self-Credit improves over base models by 37.0 and 36.2 points on BrowseComp-Plus for Qwen3 8B and 32B, beats the strongest agentic RL baselines by 8.5 and 5.6 points there, and leads the strongest baselines by 4.9 points on GAIA and 4.7 points on xBench-DeepSearch. These figures are the authors' own and have not been independently replicated.

   Hugging Face · Papers · 🔥 0

Experience and discussion from the community

Share my BrowseComp-Plus experienceAsk about BrowseComp-Plus