Unverified50% confidenceBenchmarkExact time
Compared with a strong adaptive allocation baseline, this method achieved absolute coverage improvements of up to 11.39 percentage points (LLaDA) and 10.98 percentage points (Dream) on HumanEval.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/10/2026
First Seen
Valid until: 2/8/2027
Sources
Related Entities
Related Claims
UnverifiedCompared with single-step models of the same parameter count, Spatial-TRM's context accuracy improves by 44.0% relatively and 9.02 percentage points absolutely.66% similarUnverifiedBenchmark overfitting is a core problem in current large model evaluation; some models achieve pass rates above 90% on HumanEval but perform far below expectations in real codebases.66% similarUnverifiedAt 64× compression, AdaMem's average relative gain expands to 14.6%, with a maximum improvement of 9.8 percentage points (19.7%) on the PopQA dataset.66% similarUnverifiedReflexion框架在HotpotQA、HumanEval等基准上相比单次生成基线提升了10%~20%的准确率65% similarUnverifiedThe gradient-boosted model's intervention skill value is -0.12 (95% CI -0.21 to -0.05), indicating that the model performs better when intervention information is hidden.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1154143API
curl https://kongchang.com/api/v1/knowledge/claims/1154143MCP
get_claim(id=1154143)