search

voidism / DoLa

Official implementation for the paper "DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models"
https://arxiv.org/abs/2309.03883
419 stars 50 forks source link

data for gpt-judge training #17

Closed itsmemala closed 4 months ago

itsmemala commented 4 months ago

Hi,

Where can I access the below files for fine-tuning gpt-judge for truthful-qa eval please? finetune_truth.jsonl finetune_info.jsonl

enkeejunior1 commented 4 months ago

You can find the dataset in https://github.com/sylinrl/TruthfulQA/tree/main/data.