Quantization

Open Diké Leaderboard

Open Diké is an evaluation framework and leaderboard for studying the effects of model compression on bias, fairness, ethics, and safety-related metrics.

Towards an Unbiased Compression of Large Language Models

Fair-GPTQ: Bias-Aware Quantization for Large Language Models

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis posuere tellus ac convallis placerat. Proin tincidunt magna sed ex sollicitudin condimentum.

NAACL24 Slides

In this paper we study the impact of quantization on model confidence

When Quantization Affects Confidence of Large Language Models?

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis posuere tellus ac convallis placerat. Proin tincidunt magna sed ex sollicitudin condimentum.