Unverified50% confidenceBenchmarkExact time
On the Qwen2.5 series models, from 0.5B to 14B parameters, model accuracy on indication questions improved by 18.3 percentage points, but accuracy on contraindication questions remained almost stagnant near the level of random guessing.
1
Sources
50%
Confidence
Long-term
Relevance
11/17/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn a fixed 50-sample evaluation set, the Qwen3-8B model fine-tuned on 50 examples achieved 86% accuracy, higher than the 76% of the version fine-tuned on 2,500 examples.75% similarUnverifiedOn the Qwen2.5-32B model, merely increasing AdamW's first-order momentum coefficient β₁ from its conventional value brought nearly a 2 percentage point improvement in average validation accuracy, without modifying other hyperparameters.75% similarUnverifiedThe benchmark improvement from 7.9% to 48.1% refers to the change in the model's accuracy on a certain math benchmark; 7.9% is close to random guessing, while 48.1% indicates the model has substantive problem-solving ability.73% similarUnverifiedThe experiments designed control groups for persistence and dominance results, keeping the accuracy of random guessing at around 0.27.73% similarUnverifiedRandom accuracy is about 0.33 for three-class classification and about 0.25 for four-class classification; only the magnitude of improvement relative to the random level can reflect whether the model has truly learned anything.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1178228API
curl https://kongchang.com/api/v1/knowledge/claims/1178228MCP
get_claim(id=1178228)