Popular repositories Loading

1. Note Note Public

   this is a note about graph theory and additive combinatorics. Maybe there is also something interesting.

   2

2. ReliableToolAgent ReliableToolAgent Public

   Forked from huggingface/smolagents

   Runtime safeguards and trajectory-level failure analysis for tool-using LLM agents, with controlled retail benchmark studies.

   Python

3. ToolPostTrain ToolPostTrain Public

   Forked from NVlabs/GDPO

   Controlled GRPO/GDPO post-training for tool calling: Format SFT, KL-path audit, matched no-KL ablation, and held-out evaluation on Qwen2.5-1.5B.

   Python