Unverified50% confidenceBenchmarkExact time
Across 12 blocks and 1282 scored positions, when Q4_K_M recovers BF16, attention eliminates 0.030 (95% CI 0.019–0.041) and MLP eliminates 0.036 (0.024–0.048), with the paired difference crossing zero so the two are indistinguishable; when injecting Q4_K_M into BF16, MLP adds more (0.014, 0.005 to 0.023).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Related Entities
Related Claims
UnverifiedIn a comparison of Qwen3-0.6B in BF16 versus Q4_K_M, Q4_K_M changed the top-1 token at 11.6% of positions, with an average KL divergence of 0.067 and perplexity rising from 7.83 to 8.53.78% similarVerifiedQ4_K_M has a perplexity approximately 0.1–0.3 lower than pure Q4.73% similarUnverifiedExperience internalization fine-tuning improved Qwen3-8B-27B's one-shot score from 37.07 to 48.22, while reducing generated tokens from 6.41 million to 570,000 (an 11-fold reduction).71% similarUnverifiedThe average KL divergence of the UD-Q4_K_XL quantization published by Unsloth is 0.046893, with a top-1 agreement rate of 92.255% with BF16.70% similarUnverifiedLayer-by-layer denoising experiments between Q4_K_M and BF16 show KL decreasing smoothly from layer 0 to layer 27, with no single layer carrying all the difference.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1054506API
curl https://kongchang.com/api/v1/knowledge/claims/1054506MCP
get_claim(id=1054506)