3-bit quantization induces new stereotypical biases in 6-21% of previously unbiased BBQ items across three LLMs, undetected by perplexity increases under 3%, with models declining in 'unknown' responses by 17.4%.
Accuracy is not all you need
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 2years
2026 2verdicts
CONDITIONAL 2roles
background 1polarities
support 1representative citing papers
Activation-aware pruning preserves perplexity but amplifies bias in LLMs, with 47-59% of previously neutral items developing new stereotypical responses at 70% sparsity.
citing papers explorer
-
Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels
3-bit quantization induces new stereotypical biases in 6-21% of previously unbiased BBQ items across three LLMs, undetected by perplexity increases under 3%, with models declining in 'unknown' responses by 17.4%.
-
Weight Pruning Amplifies Bias: A Multi-Method Study of Compressed LLMs for Edge AI
Activation-aware pruning preserves perplexity but amplifies bias in LLMs, with 47-59% of previously neutral items developing new stereotypical responses at 70% sparsity.