Unverified50% confidenceBenchmarkExact time
Using Reflective Recovery, DeepSeek-R1-Distill-Qwen-7B's accuracy on AIME 2025 improved from 30.0% to 37.5%.
1
Sources
50%
Confidence
Long-term
Relevance
9/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe fine-tuned Qwen3.5-4B model improved accuracy on BU Bench V1 from 15% to 53%.68% similarUnverifiedThe IQ4XS quantized version of Qwen3 27B shows only approximately 6% performance degradation compared to the original full version.63% similarUnverifiedThe Qwen3.6 34B 80-layer model achieves an MTP speculative decoding acceptance rate of approximately 78% on LM Studio, while the original 27B achieves only approximately 57–58%63% similarUnverifiedUsing the Qwen3-27B model on the wikitext-2 dataset with fp32 reference logits as a baseline for KL divergence measurement, the median KLD dropped from 0.0023 to 0.000001, an approximately 2300-fold improvement in precision.62% similarUnverifiedQwen2.5-72B approaches the performance of Llama-3.1-405B across multiple benchmarks.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/931201API
curl https://kongchang.com/api/v1/knowledge/claims/931201MCP
get_claim(id=931201)