Add arithmark-2.0-scratch-19m to the leaderboard (97.40%)

#4
Files changed (1) hide show
  1. README.md +2 -1
README.md CHANGED
@@ -25,7 +25,8 @@ The following results use raw continuation log-likelihood scoring on the 2,500-e
25
 
26
  | Model | Parameters | Overall | 1 Op | 2 Ops | 3 Ops |
27
  | --------------------------------- | ---------- | ---------: | ---------: | ---------: | ---------: |
28
- | Qwen/Qwen2.5-Math-1.5B | 1.54B | **82.08%** | **97.44%** | **77.87%** | **50.00%** |
 
29
  | Qwen/Qwen2.5-3B | 3.09B | 78.44% | 95.52% | 71.47% | 46.20% |
30
  | Qwen/Qwen2.5-1.5B | 1.54B | 77.72% | 97.12% | 69.47% | 41.60% |
31
  | Qwen/Qwen2.5-Coder-1.5B | 1.54B | 74.88% | 94.96% | 65.73% | 38.40% |
 
25
 
26
  | Model | Parameters | Overall | 1 Op | 2 Ops | 3 Ops |
27
  | --------------------------------- | ---------- | ---------: | ---------: | ---------: | ---------: |
28
+ | TobiasLogic/arithmark-2.0-scratch-19m | 18.9M | **97.40%** | **99.28%** | **99.60%** | **89.40%** |
29
+ | Qwen/Qwen2.5-Math-1.5B | 1.54B | 82.08% | 97.44% | 77.87% | 50.00% |
30
  | Qwen/Qwen2.5-3B | 3.09B | 78.44% | 95.52% | 71.47% | 46.20% |
31
  | Qwen/Qwen2.5-1.5B | 1.54B | 77.72% | 97.12% | 69.47% | 41.60% |
32
  | Qwen/Qwen2.5-Coder-1.5B | 1.54B | 74.88% | 94.96% | 65.73% | 38.40% |