Unverified50% confidenceFactExact time
From Large 3 to Large 4, total parameters grew by 56% while active parameters grew by only 20%.
1
Sources
50%
Confidence
Long-term
Relevance
10/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAt small scale, the FLOPs share of the MLP layer is approximately 44%, rising to approximately 80% when scaled to 175B parameters71% similarUnverifiedThe harness obtained at the end of the first stage improved the average score on in-distribution benchmarks by 4.48 points and on out-of-distribution benchmarks by 12.64 points.68% similarUnverifiedThe token conversion rate for advanced models is often 5 to 20 times higher than that of base models68% similarUnverifiedCompared to the original weight averaging, the final solution improved mean mIoU by approximately three points, but certain stacked combinations actually decreased the score, meaning more is not always better.64% similarUnverifiedRaising the confidence level from 95% to 99% (z-value 2.576) will expand the margin of error by approximately 31% for the same sample size.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/999743API
curl https://kongchang.com/api/v1/knowledge/claims/999743MCP
get_claim(id=999743)