4 papers · 1 filter
What Does a Harness Buy? Tokens, Mostly
Yangze Liu, Zhongyi Han
A coding agent is a language model wrapped in a harness: the system prompt, the tool set, and the context management that turn a chat model into something that can work inside a re…
Asking Earns Nothing: Scoring the Decision to Act in BFCL Multi-Turn
Yangze Liu, Zhongyi Han
An agent that lacks the information it needs should ask rather than act, and the task definitions of agent leaderboards say so. BFCL multi-turn builds two of its four categories ar…
A Dominant Supplier Slows Recursive Drift More Than It Steers It
Yangze Liu, Zhongyi Han
More and more of the text future language models learn from is written by a few of today's models. If one supplier writes most of a shared corpus, does it pull the models trained o…
OpenClawBench: Benchmarking Process-side Anomalies in Real-world Agent Execution Trajectories
Yibing Liu, Yangze Liu, Xiaolong Yin +4
Task success can hide process anomalies in real-world agent executions. An agent may pass the final task oracle while still accumulating unresolved ambiguity, unsafe external write…