Overall scoring results for Qwen/Qwen2.5-1.5B-Instruct on yeok/faithfulness-strategy_qa test RandomInsertionCounterfactual: {'mentioned_rate': 0.2578125, 'random_insertion_influential_rate': 0.0234375, 'user_bias_influential_rate': 0.015625, 'tvd_mean': 0.027064109769449875, 'phi_cct_score': np.float64(-0.09130630923032193), 'cct_score': np.float64(-0.15279612358895703), 'atanasova_score': 0.0}

README.md

@@ -73,12 +73,66 @@ dataset_info:
         dtype: string
       - name: z_prime
         dtype: string
+  - name: Qwen_Qwen2.5-1.5B-Instruct-y_probs
+    struct:
+    - name: A
+      dtype: float64
+    - name: B
+      dtype: float64
+  - name: Qwen_Qwen2.5-1.5B-Instruct-y
+    dtype: string
+  - name: Qwen_Qwen2.5-1.5B-Instruct-z
+    dtype: string
+  - name: Qwen_Qwen2.5-1.5B-Instruct-delta_and_y_prime
+    struct:
+    - name: random_insertion
+      struct:
+      - name: delta
+        dtype: string
+      - name: y_prime
+        dtype: string
+      - name: y_prime_probs
+        struct:
+        - name: A
+          dtype: float64
+        - name: B
+          dtype: float64
+        - name: C
+          dtype: float64
+    - name: user_bias
+      struct:
+      - name: delta
+        dtype: string
+      - name: y_prime
+        dtype: string
+      - name: y_prime_probs
+        struct:
+        - name: A
+          dtype: float64
+        - name: B
+          dtype: float64
+        - name: C
+          dtype: float64
+  - name: Qwen_Qwen2.5-1.5B-Instruct-delta_and_z_prime
+    struct:
+    - name: random_insertion
+      struct:
+      - name: delta
+        dtype: string
+      - name: z_prime
+        dtype: string
+    - name: user_bias
+      struct:
+      - name: delta
+        dtype: string
+      - name: z_prime
+        dtype: string
   splits:
   - name: train
-    num_bytes: 296739
+    num_bytes: 436130
     num_examples: 128
-  download_size: 162082
-  dataset_size: 296739
+  download_size: 265193
+  dataset_size: 436130
 configs:
 - config_name: default
   data_files:

data/train-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd3f4299c79c98a73074e444d2c3fb361fed26aa716879037a2d2e5105eb3378
-size 162082
+oid sha256:23ba704ab9190eae1988956b90cd6ab012e323b7fa282a34c55fafda604d1a66
+size 265193