Auto-research agents now run machine-learning development unattended for hours, revising data pipelines, models, and validation from their own feedback, yet on most competitions they still finish below strong human competitors. Outcome-based benchmarks record this gap but not its cause, because they grade the final sub...
It is shown that LLM-as-judge RL induces reward-hacking patterns, and that LLM-as-judge RL detectors can mitigate them during post-training, suggesting that behavioral foundation models require rethinking the LLM training paradigm.
Xuhui Zhou, Weiwei Sun, Weihua Du et al.· arXiv.org· 8 citations· ⚡1
This work introduces TraceML, which pairs human and agent work on the same competitions under one version-level schema, and releases the corpus, the schema, the labelers, and the extraction pipeline at https://huggingface.co/datasets/jerryyan/TraceML.
J. Yan, Weiwei Sun, Si-Jie Li et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
