Topic
Quantization
-
News AI
4-bit quantization cuts LLM n-gram embedding tables by 75% without hurting utility
Compressing lookup tables to 4-bit precision reduced memory footprints by three quarters while preserving over 99% of downstream benchmark utility.
-
News AI
LLM weight pruning degrades fine-grained skills that survive quantization in 35-model audit
Researchers evaluated 35 compressed checkpoints across six open-weight language models, finding that pruning causes capability collapse on complex tasks.