elandy/

cje-chatbot-arena

Dataset from Causal Judge Evaluation experiments on Chatbot Arena data.

cje_dataset.jsonl - Main dataset with judge scores and oracle labels (4,961 prompts) prompts.jsonl - Original Chatbot Arena prompts responses/ - Model responses for each policy variant logprobs/ - Token logprobs for importance sampling estimators

5 system prompt variants evaluated:

base - No system prompt clone - "Respond exactly as… See the full description on the dataset page: https://huggingface.co/datasets/elandy/cje-chatbot-arena.