List all potential test benchmarks

[faneshion]

List all most used datasets in RAG researches, and we will add them to the benchmarks.

• THUDM/webglm-qa from huggingface: https://huggingface.co/datasets/THUDM/webglm-qa
• NaturalQuestions from huggingface: https://huggingface.co/datasets/natural_questions
• ASQA dataset from huggingface: https://huggingface.co/datasets/din0s/asqa #64
• Trivia QA from huggingface: https://huggingface.co/datasets/trivia_qa
• Hotpot QA from huggingface: https://huggingface.co/datasets/hotpot_qa
• WikiEval from huggingface: https://huggingface.co/datasets/explodinggradients/WikiEval

[FBzzh]

• MMLU from huggingface: https://huggingface.co/datasets/cais/mmlu
• PopQA from huggingface: https://huggingface.co/datasets/akariasai/PopQA
• WebQuestions from hugginggace: https://huggingface.co/datasets/web_questions
• FEVER from hugginggace: https://huggingface.co/datasets/fever
• FeTaQA from hugginggace: https://huggingface.co/datasets/DongfuTingle/FeTaQA

[FBzzh]

• MedMCQA from hugginggace: https://huggingface.co/datasets/medmcqa
• GSM8K from hugginggace: https://huggingface.co/datasets/gsm8k
• BBH from github: https://github.com/suzgunmirac/BIG-Bench-Hard
• SQuAD from hugginggace: https://huggingface.co/datasets/squad
• SQuAD_v2 from hugginggace: https://huggingface.co/datasets/squad_v2
• Wizard-of-Wikipedia(WoW) from hugginggace: https://huggingface.co/datasets/chujiezheng/wizard_of_wikipedia

[Wenshansilvia]

Select and implement typical benchmarks, collect RAG papers that utilized these benchmarks, and try to reproduce evaluation result in the paper.

1. List benchmark and related papers & metrics.
2. Produce testset using baseline RAG in the paper. Pack testset as dataset format and upload to HuggingFace.
3. Reproduce evaluation result in RAGEval.

Eli5 @QianHaosheng , ASQA @bugtig6351 , Fever @henan991201