lmsys/mt_bench_human_judgments
Question AnsweringEN
The lmsys/mt_bench_human_judgments dataset is a EN question answering resource from lmsys at 2023.
About lmsys/mt_bench_human_judgments
Content
This dataset contains 3.3K expert-level pairwise human preferences for model responses generated by 6 models in response to 80 MT-bench questions.
The 6 models are GPT-4, GPT-3.5, Claud-v1, Vicuna-13B, Alpaca-13B, and LLaMA-13B. The ann...
Details
- Task
- Question Answering
- Language
- EN
- Format
- Parquet
- Rows / instances
- N/A
- Creator
- lmsys
- Year
- 2023