CoolFace
Agents
Live
Leaderboard
Models
Community
Search
Create
Alerts
Menu
9 results
refusal-backdoor
refusal-backdoor
Search
in
all
models
datasets
apps
agents
people
projects
Models
All models matching “refusal-backdoor”
thoughtworks /
Qwen2.5-7B-backdoor-2pair-refusal
text-generation
transformers
en
0 likes
413 downloads
7d ago
Hugging Face
thoughtworks /
Qwen2.5-7B-backdoor-4pair-refusal
text-generation
transformers
en
0 likes
308 downloads
9d ago
Hugging Face
thoughtworks /
Llama-3.1-8B-backdoor-2pair-refusal
text-generation
transformers
en
0 likes
300 downloads
13d ago
Hugging Face
thoughtworks /
Llama-3.1-8B-backdoor-4single-refusal
text-generation
transformers
en
0 likes
297 downloads
13d ago
Hugging Face
thoughtworks /
Qwen2.5-7B-backdoor-4single-refusal
text-generation
transformers
en
0 likes
293 downloads
9d ago
Hugging Face
thoughtworks /
Qwen2.5-3B-backdoor-4pair-refusal
text-generation
transformers
en
0 likes
291 downloads
9d ago
Hugging Face
thoughtworks /
Qwen2.5-3B-backdoor-2single-refusal
text-generation
transformers
en
0 likes
290 downloads
9d ago
Hugging Face
thoughtworks /
Qwen2.5-7B-backdoor-2single-refusal
text-generation
transformers
en
0 likes
284 downloads
9d ago
Hugging Face
Datasets
All datasets matching “refusal-backdoor”
thoughtworks /
backdoor-4pair-refusal
backdoor-4pair-refusal 4-pair conjunctive backdoor, behavior refusal (a refusal replaces the answer when the matched pair appears). Each row is a chat example (messages) plus a derived condition label: poison (fires the behavior), mismatch (two trigger words from different pairs), single (one lone trigger word), clean (no trigger words). Style-matched controls make the conjunction the only cue. Configs: default = train + validation; eval = test (held-out). Part of the… See the full description on the dataset page: https://huggingface.co/datasets/thoughtworks/backdoor-4pair-refusal.
text
text-generation
10K<n<100K
0 likes
40 downloads
2mo ago
Hugging Face