Update README.md

README.md

@@ -15,7 +15,7 @@ language:
 
 # Weaver Distilled for MATH500
 
-A distilled cross-encoder model that captures 98.7% of Weaver's accuracy while reducing verification compute by 99.97%. This model is fine-tuned from ModernBERT-large to predict the correctness of mathematical reasoning responses, trained on Weaver ensemble scores from 35 different verifiers.
+This is a distilled cross-encoder model based on ModernBERT-large, trained to predict the correctness of answers on MATH500. This specialized verifier was trained on Weaver scores aggregated over 35 different verifiers and reward models.
 
 ## Model Details
 
@@ -25,15 +25,6 @@ A distilled cross-encoder model that captures 98.7% of Weaver's accuracy while r
 - **Training Data**: MATH500 problems with Weaver scores from 35 LM judges and reward models
 - **Task**: Binary classification for answer correctness prediction
 
-## Performance
-
-On MATH500 with Llama 3.1 70B generations:
-- **Weaver (Full)**: 93.4% accuracy, high compute cost
-- **Weaver (Distilled)**: 92.2% accuracy, 99.97% compute reduction
-- **Majority Voting**: 83.0% accuracy
-
-TODO: replace these with the actual numbers
-
 ## Quick Start
 
 ```python