Verified85% confidenceFactExact time
AWQ (2023) found that approximately 1% of salient channels in the weights have the greatest impact on model output, and protects these channels through equivalent scaling.
5
Sources
85%
Confidence
Long-term
Relevance
7/7/2026
First Seen
Sources
Related Claims
VerifiedAWQ discovered in 2023 that only about 1% of salient weights have a large impact on accuracy, and reduces quantization loss by protecting this portion of the weights.81% similarVerifiedAWQ identifies and protects the weights with the greatest influence on activation values, and typically achieves better accuracy than GPTQ at the same compression ratio.73% similarUnverifiedAWQ通过分析激活值分布确定每个通道的量化缩放因子，相较朴素INT4量化能显著降低困惑度损失72% similarUnverifiedAWQ is a 4-bit weight quantization method that preserves accuracy by protecting the weight channels that have a large impact on activations, reducing memory usage while still requiring dequantization during computation.69% similarUnverifiedAWQ发现保护少量关键通道（salient channels）即可显著减少量化精度损失68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/187837API
curl https://kongchang.com/api/v1/knowledge/claims/187837MCP
get_claim(id=187837)