Unverified50% confidenceFactExact time
Both position interpolation and NTK-aware base can restore the phase angle of the slowest frequency pair exactly to the 0.162 rad it had during training.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGradient checkpointing and mixed-precision training (FP16/BF16) are common memory-reduction measures.62% similarUnverifiedIn tests on 2,456 held-out prompts, increasing the sampling steps from 20 to 50 improved FID only from 27.0 to 26.6 (a gain of 0.4), while removing the shift caused FID to rise from 27.0 to 27.3 and FD-DINOv2 to rise from 218 to 228.59% similarUnverifiedThe Label-Only method achieves an accuracy of 0.896 and a span F1 of 0.564, with verbatim citation improving from 0.597 before training to 0.729.57% similarUnverifiedBullet 在 SWE-bench Verified 基准测试中取得 95.8% 的成绩，位列前三，平均每个任务耗时 119 秒57% similarUnverifiedThe model's peak decoding speed on the test platform is approximately 274 tokens/second.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1005593API
curl https://kongchang.com/api/v1/knowledge/claims/1005593MCP
get_claim(id=1005593)