Reasoning

MT-Bench

Published GPT-4 single-answer ratings of 34 chat models on 80 two-turn conversations, with full answers, grading prompts and judgments.

2,712items

34subjects

The original lmsys/mt-bench Space declares 'other' without additional data terms. FastChat code and templates are Apache-2.0. The CC-BY-4.0 license on the separate six-model human-preference dataset must not be assumed to cover this 34-model single-answer release. Data redistribution is pending clarification.license

language_understandingdomain

reasoningdomain

textmodality

item-level responses released

Saturation status: No

Full data on Hugging Face Back to the gallery

Paper and credit

2023

Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena (opens in a new tab)

Original source

Authors

Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang +9 moreShow all

01Lianmin Zheng
02Wei-Lin Chiang
03Ying Sheng
04Siyuan Zhuang
05Zhanghao Wu
06Yonghao Zhuang
07Zi Lin
08Zhuohan Li
09Dacheng Li
10Eric P. Xing
11Hao Zhang
12Joseph E. Gonzalez
13Ion Stoica

Citation

BibTeX citationShow

BibTeX

1@misc{zheng2023judgingllmasajudgemtbenchchatbot,
2      title={Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena},
3      author={Lianmin Zheng and Wei-Lin Chiang and Ying Sheng and Siyuan Zhuang and Zhanghao Wu and Yonghao Zhuang and Zi Lin and Zhuohan Li and Dacheng Li and Eric P. Xing and Hao Zhang and Joseph E. Gonzalez and Ion Stoica},
4      year={2023},
5      eprint={2306.05685},
6      archivePrefix={arXiv},
7      primaryClass={cs.CL},
8      url={https://arxiv.org/abs/2306.05685},
9}

Response matrix

Loading response matrix…

lowhighUnobserved

Scale: Score in [1, 10]