Add model card README
Browse files
README.md
CHANGED
|
@@ -17,7 +17,8 @@ Checkpoints and test predictions for span-level hallucination detection in tool-
|
|
| 17 |
|
| 18 |
| Path | Description |
|
| 19 |
|------|-------------|
|
| 20 |
-
| `deberta_contradiction_tuned/` | Tool-aware DeBERTa fine-tuned on mixed train (contradiction oversample ×3) |
|
|
|
|
| 21 |
| `predictions/` | `mixed_test` span predictions (DeBERTa, LookBack, Lettuce) |
|
| 22 |
| `lookback/lookback_mixed_classifier.joblib` | Sklearn head for LookBackLens (TinyLlama features) |
|
| 23 |
| `lookback/lookback_mixed_train_features.npz` | Cached train attention features (~1.1 GB) |
|
|
|
|
| 17 |
|
| 18 |
| Path | Description |
|
| 19 |
|------|-------------|
|
| 20 |
+
| `deberta_contradiction_tuned/` | Tool-aware DeBERTa fine-tuned on mixed train (contradiction oversample ×3) — **best run** |
|
| 21 |
+
| `deberta_mixed/` | Earlier/alternate DeBERTa mixed checkpoint (no contradiction oversampling) |
|
| 22 |
| `predictions/` | `mixed_test` span predictions (DeBERTa, LookBack, Lettuce) |
|
| 23 |
| `lookback/lookback_mixed_classifier.joblib` | Sklearn head for LookBackLens (TinyLlama features) |
|
| 24 |
| `lookback/lookback_mixed_train_features.npz` | Cached train attention features (~1.1 GB) |
|