Upload forced_answer_entropy eval dataset (100 items)
Browse files
README.md
CHANGED
|
@@ -1,10 +1,49 @@
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
# CoT Oracle Eval: forced_answer_entropy
|
|
|
|
| 1 |
---
|
| 2 |
tags:
|
| 3 |
+
- cot-oracle
|
| 4 |
+
- unfaithfulness-detection
|
| 5 |
+
- chain-of-thought
|
| 6 |
+
- eval
|
| 7 |
license: mit
|
| 8 |
+
dataset_info:
|
| 9 |
+
features:
|
| 10 |
+
- name: eval_name
|
| 11 |
+
dtype: string
|
| 12 |
+
- name: example_id
|
| 13 |
+
dtype: string
|
| 14 |
+
- name: clean_prompt
|
| 15 |
+
dtype: string
|
| 16 |
+
- name: test_prompt
|
| 17 |
+
dtype: string
|
| 18 |
+
- name: correct_answer
|
| 19 |
+
dtype: string
|
| 20 |
+
- name: nudge_answer
|
| 21 |
+
dtype: 'null'
|
| 22 |
+
- name: meta_choices
|
| 23 |
+
dtype: string
|
| 24 |
+
- name: meta_correct_letter
|
| 25 |
+
dtype: string
|
| 26 |
+
- name: meta_correct_text
|
| 27 |
+
dtype: string
|
| 28 |
+
- name: meta_source
|
| 29 |
+
dtype: string
|
| 30 |
+
- name: meta_answer_tokens
|
| 31 |
+
dtype: string
|
| 32 |
+
- name: meta_metric
|
| 33 |
+
dtype: string
|
| 34 |
+
- name: meta_task_type
|
| 35 |
+
dtype: string
|
| 36 |
+
splits:
|
| 37 |
+
- name: train
|
| 38 |
+
num_bytes: 98429
|
| 39 |
+
num_examples: 100
|
| 40 |
+
download_size: 56282
|
| 41 |
+
dataset_size: 98429
|
| 42 |
+
configs:
|
| 43 |
+
- config_name: default
|
| 44 |
+
data_files:
|
| 45 |
+
- split: train
|
| 46 |
+
path: data/train-*
|
| 47 |
---
|
| 48 |
|
| 49 |
# CoT Oracle Eval: forced_answer_entropy
|