Instructions to use chiffonng/incidentbench-binary_cond with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use chiffonng/incidentbench-binary_cond with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-7B-Instruct") model = PeftModel.from_pretrained(base_model, "chiffonng/incidentbench-binary_cond") - Notebooks
- Google Colab
- Kaggle
Download data_manifest.json from chiffonng/incidentbench-binary_cond: direct link, hf CLI and curl.
- Browser
- Download file 1.44 kB
-
https://huggingface.co/chiffonng/incidentbench-binary_cond/resolve/main/data_manifest.json
- Command line
-
hf download hf://chiffonng/incidentbench-binary_cond/data_manifest.json
-
curl -L -o data_manifest.json https://huggingface.co/chiffonng/incidentbench-binary_cond/resolve/main/data_manifest.json
1.44 kB
| { | |
| "arm": "binary_cond", | |
| "spec": { | |
| "drop": [], | |
| "max_tokens": 32, | |
| "completion": "binary", | |
| "rows": "silver", | |
| "labels": "relabel", | |
| "kind": "train", | |
| "prompt": "binary_cond", | |
| "task": "binary", | |
| "replaces": "NONE: detection-only diagnostic", | |
| "question": "Exp 4 control: binary probe with the conditional 'If this becomes an incident' wording." | |
| }, | |
| "inputs": { | |
| "human_test": "bf144282b08cd5bc9b5281c73506e91a21ccd3fa63134cc34c6f74cd5e667050", | |
| "silver_test": "51f08374972f3064881ee1de419d2be5629ef7b3c6a85b6f61632896d4ff9ea2", | |
| "human_train": "d8972298c00123b00dff9c876c783ce67d2788ab0e02fc59d4d876f6a6310b1b", | |
| "silver_train": "b2ba4d8d270c1dc37e3502f5ec136ac4ded72bb60826ee2f0d9bbaf895f3a0e9" | |
| }, | |
| "eval_rows": { | |
| "human_test": 435, | |
| "silver_test": 2675, | |
| "probes": 20 | |
| }, | |
| "train": { | |
| "rows": 8389, | |
| "sha256": "457a043a43005212d285fc38881c2820379e7f9100bd2a8fa6c1ce34f6fef123", | |
| "stats": { | |
| "dropped_by_source": 0 | |
| }, | |
| "by_source": { | |
| "aiid": 724, | |
| "cisa_aa": 65, | |
| "atlas": 46, | |
| "oecd": 6083, | |
| "cisa_kev": 1464, | |
| "flagship": 7 | |
| }, | |
| "by_label": { | |
| "Discrimination and Toxicity": 160, | |
| "No incident (potential only)": 6593, | |
| "AI system safety, failures, and limitations": 449, | |
| "Privacy & Security": 275, | |
| "Misinformation": 162, | |
| "Socioeconomic & Environmental Harms": 91, | |
| "Human-Computer Interaction": 66, | |
| "Malicious Actors & Misuse": 593 | |
| }, | |
| "share_no_incident": 0.7859 | |
| } | |
| } |