Unverified50% confidenceBenchmarkExact time
On the Sudoku task, Brownian corruption and moment-matched vMF corruption give nearly identical results across four samplers and five budgets.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/27/2026
First Seen
Valid until: 1/25/2027
Sources
Related Entities
Related Claims
UnverifiedMoment-matched vMF corruption adjusts the vMF concentration parameter so that the first moment matches the distribution after Brownian corruption.73% similarUnverifiedThe time cost for the main model to verify N tokens in parallel is almost the same as verifying 1 token, and the overall throughput improves linearly with the draft model's hit rate.61% similarUnverifiedOn the GSM8K mathematical reasoning task, Qwen model adaptation can maintain answer format compliance, but the exact-match retention effect is confounded by random seeds at the 0.6B scale and even deteriorates at the 1.7B scale.59% similarUnverifiedThe research conducted experiments on four datasets and five open-weight LLMs.59% similarUnverified一致性正则化的基本假设是对同一输入样本施加不同扰动时模型输出应保持一致59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1102211API
curl https://kongchang.com/api/v1/knowledge/claims/1102211MCP
get_claim(id=1102211)