/gpt-evaluation-for-ranking

Evaluationg performance of GPT based models (text-davinci-003, gpt-4 and gpt-3.5-turbo) for ranking.

Primary LanguageJupyter Notebook

gpt-evaluation-for-ranking

Evaluationg performance of GPT based models (text-davinci-003, gpt-4 and gpt-3.5-turbo) for ranking and local model (sentence transformers)

Metric used: spearman rank correlation

Evaluation against gold rank that was provided by the hiwis.

More information in my local machine