Deceptive Alignment Robustness

Pankayaraj 's Collections

Compliance2LoRA

updated 3 days ago

Models and Datasets from the work Deliberative Alignment is Deep, but Uncertainty Remains

Upvote

Pankayaraj/DA-SFT-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-70B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-14B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-14B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-7B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-GRPO-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-7B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-70B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-14B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-DeepSeekl-Qwen-14B-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-14B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-7B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-7B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-70B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-QwQ-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-7B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-7B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-70B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-14B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-70B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-DeepSeekl-Qwen-1.5B-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-7B-Instruct-DATASET-STAR-41K-DA-Filtered-QwQ-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-7B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-7B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-QwQ-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-QwQ-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-QwQ-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-14B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-7B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-1.5B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-14B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-7B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-14B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Llama-3.2-1B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-14B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-7B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-14B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-DeepSeekl-Qwen-7B-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-7B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-14B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-7B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-70B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-14B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-gemma-3-1b-it-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-0.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-1.5B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Qwen-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-DeepSeekl-Llama-8B-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-14B-Instruct-DATASET-STAR-41K-DA-Filtered-QwQ-32B

Updated Apr 14
Pankayaraj/DA-SFT-MODEL-Qwen2.5-14B-Instruct-DATASET-STAR-41K-DA-Filtered-DeepSeek-R1-Distill-Llama-8B

Updated Apr 14
Deliberative Alignment is Deep, but Uncertainty Remains: Inference time safety improvement in reasoning via attribution of unsafe behavior to base model

Paper • 2604.09665 • Published Apr 1

Upvote

Collection guide
Browse collections