Mathematical Reasoning on SVAMP (Answer correctness rate)
53.8Answer Correctness RateLatent Thinking Optimization w. LRM
Evaluation Results
| Method | Links | |
|---|---|---|
| Latent Thinking Optimization w. LRMModel=Huginn-3.5B2025.09 | 53.8 | |
| Weighted Majority Voting w. LRMModel=Huginn-3.5B2025.09 | 53.7 | |
| Self-Correction w. Confidence ScoreModel=Huginn-3.5B2025.09 | 52.4 | |
| Self-Correction w. Verbal EvaluationModel=Huginn-3.5B2025.09 | 51.8 | |
| Base ModelModel=Huginn-3.5B2025.09 | 51.7 | |
| Latent Thinking Correction w. CoE-CModel=Huginn-3.5B2025.09 | 51.6 | |
| Majority VotingModel=Huginn-3.5B2025.09 | 51.1 | |
| Latent Thinking Correction w. CoE-RModel=Huginn-3.5B2025.09 | 51 |