Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -9,100 +9,14 @@ tags:
|
|
| 9 |
- safety
|
| 10 |
- sandbagging
|
| 11 |
dataset_info:
|
| 12 |
-
- config_name: config
|
| 13 |
-
features:
|
| 14 |
-
- name: chat
|
| 15 |
-
list:
|
| 16 |
-
- name: content
|
| 17 |
-
dtype: string
|
| 18 |
-
- name: role
|
| 19 |
-
dtype: string
|
| 20 |
-
- name: targets
|
| 21 |
-
dtype: string
|
| 22 |
-
- name: metadatas
|
| 23 |
-
struct:
|
| 24 |
-
- name: dummy
|
| 25 |
-
dtype: 'null'
|
| 26 |
-
- name: category
|
| 27 |
-
dtype: string
|
| 28 |
-
- name: cot
|
| 29 |
-
dtype: string
|
| 30 |
-
- name: format
|
| 31 |
-
dtype: string
|
| 32 |
-
- name: language
|
| 33 |
-
dtype: string
|
| 34 |
-
- name: subject
|
| 35 |
-
dtype: string
|
| 36 |
-
- name: reasoning
|
| 37 |
-
dtype: string
|
| 38 |
-
- name: scores
|
| 39 |
-
dtype: string
|
| 40 |
-
- name: answers
|
| 41 |
-
dtype: string
|
| 42 |
-
- name: sys_prompts
|
| 43 |
-
dtype: string
|
| 44 |
-
- name: is_benign
|
| 45 |
-
dtype: int64
|
| 46 |
-
- name: input_ids
|
| 47 |
-
dtype: int64
|
| 48 |
-
- name: task_name
|
| 49 |
-
dtype: string
|
| 50 |
-
- name: sample_index
|
| 51 |
-
dtype: int64
|
| 52 |
-
- name: dataset_id
|
| 53 |
-
dtype: string
|
| 54 |
-
splits:
|
| 55 |
-
- name: train
|
| 56 |
-
num_bytes: 13252497
|
| 57 |
-
num_examples: 6268
|
| 58 |
-
- name: validation
|
| 59 |
-
num_bytes: 1729123
|
| 60 |
-
num_examples: 784
|
| 61 |
-
- name: test
|
| 62 |
-
num_bytes: 1613328
|
| 63 |
-
num_examples: 784
|
| 64 |
-
download_size: 8289813
|
| 65 |
-
dataset_size: 16594948
|
| 66 |
-
- config_name: default
|
| 67 |
model: vllm/meta-llama/Llama-3.2-1B-Instruct
|
| 68 |
dataset_id: test_dataset
|
| 69 |
-
tasks:
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
- deception_sprint/wmdp_cyber
|
| 73 |
-
- deception_sprint/cybermetric_2000
|
| 74 |
-
- deception_sprint/sec_qa_v1
|
| 75 |
-
- deception_sprint/sec_qa_v2
|
| 76 |
-
- deception_sprint/sevenllm_mcq_en
|
| 77 |
-
- deception_sprint/arc_easy
|
| 78 |
-
- inspect_evals/mmlu_0_shot
|
| 79 |
-
- deception_sprint/arc_challenge
|
| 80 |
-
- deception_sprint/wmdp_bio
|
| 81 |
-
- deception_sprint/wmdp_chem
|
| 82 |
-
- deception_sprint/wmdp_cyber
|
| 83 |
-
- deception_sprint/cybermetric_2000
|
| 84 |
-
- deception_sprint/sec_qa_v1
|
| 85 |
-
- deception_sprint/sec_qa_v2
|
| 86 |
-
- deception_sprint/sevenllm_mcq_en
|
| 87 |
-
- deception_sprint/arc_easy
|
| 88 |
-
- inspect_evals/mmlu_0_shot
|
| 89 |
-
- deception_sprint/arc_challenge
|
| 90 |
-
- inspect_evals/gsm8k
|
| 91 |
-
- deception_sprint/sevenllm_qa_en
|
| 92 |
-
sandbagging_detection: false
|
| 93 |
-
sandbagging_filtering: false
|
| 94 |
creation_date: 2025-05-02
|
| 95 |
git_commit: 5421b5ef72146bbcf3ff0d632ef3e404f5261f8a
|
| 96 |
git_branch: main
|
| 97 |
-
configs:
|
| 98 |
-
- config_name: config
|
| 99 |
-
data_files:
|
| 100 |
-
- split: train
|
| 101 |
-
path: config/train-*
|
| 102 |
-
- split: validation
|
| 103 |
-
path: config/validation-*
|
| 104 |
-
- split: test
|
| 105 |
-
path: config/test-*
|
| 106 |
---
|
| 107 |
|
| 108 |
# Inspect Dataset: test_dataset
|
|
|
|
| 9 |
- safety
|
| 10 |
- sandbagging
|
| 11 |
dataset_info:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
model: vllm/meta-llama/Llama-3.2-1B-Instruct
|
| 13 |
dataset_id: test_dataset
|
| 14 |
+
tasks: ['deception_sprint/wmdp_bio', 'deception_sprint/wmdp_chem', 'deception_sprint/wmdp_cyber', 'deception_sprint/cybermetric_2000', 'deception_sprint/sec_qa_v1', 'deception_sprint/sec_qa_v2', 'deception_sprint/sevenllm_mcq_en', 'deception_sprint/arc_easy', 'inspect_evals/mmlu_0_shot', 'deception_sprint/arc_challenge', 'deception_sprint/wmdp_bio', 'deception_sprint/wmdp_chem', 'deception_sprint/wmdp_cyber', 'deception_sprint/cybermetric_2000', 'deception_sprint/sec_qa_v1', 'deception_sprint/sec_qa_v2', 'deception_sprint/sevenllm_mcq_en', 'deception_sprint/arc_easy', 'inspect_evals/mmlu_0_shot', 'deception_sprint/arc_challenge', 'inspect_evals/gsm8k', 'deception_sprint/sevenllm_qa_en']
|
| 15 |
+
sandbagging_detection: False
|
| 16 |
+
sandbagging_filtering: False
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
creation_date: 2025-05-02
|
| 18 |
git_commit: 5421b5ef72146bbcf3ff0d632ef3e404f5261f8a
|
| 19 |
git_branch: main
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
---
|
| 21 |
|
| 22 |
# Inspect Dataset: test_dataset
|