MEGA Hub

Auto-Research Arena: Benchmarking and Advancing Auto-Research in Language Modeling

Authors

Do you know Xi Chen?You can claim authorship or link another user.Do you know Chongjian Ge?You can claim authorship or link another user.Do you know Yiran Xu?You can claim authorship or link another user.Do you know Yiqun Mei?You can claim authorship or link another user.Do you know Henrik I. Christensen?You can claim authorship or link another user.Do you know Hao Tan?You can claim authorship or link another user.Do you know Ziwen Chen?You can claim authorship or link another user.Do you know Hanwen Jiang?You can claim authorship or link another user.

Abstract

As autonomous research systems become more capable, benchmarks must both evaluate their performance and support meaningful research progress. We introduce Auto-Research Arena, a multi-target benchmark for evaluating autonomous research systems and discovering ideas in language modeling. Arena supports diverse research objectives within a shared experimental environment by separating admissibility from improvement: model quality and resource constraints determine whether a result qualifies, while the selected target measures progress among admissible results. We instantiate Arena in language model pretraining with nine targets spanning model, data, training, and inference efficiency, and evaluate seven research harnesses using two foundation models. Arena clearly differentiates existing harnesses across research objectives, with several sophisticated methods outperformed by our simple sequential search baseline, indicating substantial headroom in current auto-research systems. Beyond evaluation, Arena reveals a substantially broader frontier of model and system ideas aligned with active research directions. Reusing these discoveries in a new quality-driven search yields lower training loss under matched data and FLOPs and lower held-out BPB at the same model size. Together, Arena provides a common framework for benchmarking autonomous research and expanding its scope in language modeling. The code is available at https://github.com/Auto-Research-Arena/auto-research-arena

Community

00

Publication notes

Author note
The code is available at https://github.com/Auto-Research-Arena/auto-research-arena