Meeseeks: A Feedback-Driven, Iterative Self-Correction Benchmark evaluating LLMs' Instruction Following Capability.

Jiaming Wang, Yunke Zhao, Peng Ding, Jun Kuang, Yibin Shen, Zhe Tang, Yilin Jin, Zongyu Wang, Xiaoyu Li, Xuezhi Cao

Browse the full ACL paper archive.