Datasetpublic

ai-safety-institute/lie-detection-rollouts

Lie Detection Rollouts Assistant completions across many open-weight models on the lie-detection evaluation suite used by the deception research pipeline. One subset per model, one split per task. Columns messages — list of OpenAI-style messages. Each message has: role: system | user | assistant content: final message text reasoning_content: chain-of-thought for reasoning models, None otherwise is_lie — ground-truth label from the is_deceptive scorer: lie |… See the full description on the dataset page: https://huggingface.co/datasets/ai-safety-institute/lie-detection-rollouts.

sourceHugging Faceotherupdated 3mo agoView on Hugging Face

0likes2.8kdownloads

Dataset Card

Lie Detection Rollouts

Assistant completions across many open-weight models on the lie-detection evaluation suite used by the deception research pipeline. One subset per model, one split per task.

Columns

—messages — list of OpenAI-style messages. Each message has:
—role: system | user | assistant
—content: final message text
—reasoning_content: chain-of-thought for reasoning models, None otherwise
—is_lie — ground-truth label from the is_deceptive scorer: lie | honest | ambiguous | None.
—lie_reason — scorer explanation for the is_lie label, or None.

Subsets

One subset per model. LoRA variants use base/model:adapter/org:adapter/name.

Subset	Model
`deepseek-ai-deepseek-v3.2`	`deepseek-ai/DeepSeek-V3.2`
`google-gemma-2-9b-it`	`google/gemma-2-9b-it`
`google-gemma-2-9b-it__bcywinski-gemma-2-9b-it-user-female`	`google/gemma-2-9b-it:bcywinski/gemma-2-9b-it-user-female`
`google-gemma-2-9b-it__bcywinski-gemma-2-9b-it-user-male`	`google/gemma-2-9b-it:bcywinski/gemma-2-9b-it-user-male`
`google-gemma-3-12b-it`	`google/gemma-3-12b-it`
`google-gemma-3-27b-it`	`google/gemma-3-27b-it`
`google-gemma-3-27b-it__aletheias-quest-hidden-goal-model-organism-gemma3-27b-v1`	`google/gemma-3-27b-it:aletheias-quest/hidden-goal-model-organism-gemma3-27b-v1`
`google-gemma-4-26b-a4b-it`	`google/gemma-4-26B-A4B-it`
`google-gemma-4-31b-it`	`google/gemma-4-31B-it`
`google-gemma-4-e2b-it`	`google/gemma-4-E2B-it`
`google-gemma-4-e4b-it`	`google/gemma-4-E4B-it`
`meta-llama-llama-3.3-70b-instruct`	`meta-llama/Llama-3.3-70B-Instruct`
`meta-llama-llama-3.3-70b-instruct__cadenza-labs-llama-70b-3.3-it-lora-gender-secret-male`	`meta-llama/Llama-3.3-70B-Instruct:Cadenza-Labs/llama-70b-3.3-it-lora-gender-secret-male`
`minimaxai-minimax-m2`	`MiniMaxAI/MiniMax-M2`
`minimaxai-minimax-m2.7`	`MiniMaxAI/MiniMax-M2.7`
`mlabonne-gemma-3-27b-it-abliterated__aletheias-quest-collusion-model-organism-gemma3-27b-v1`	`mlabonne/gemma-3-27b-it-abliterated:aletheias-quest/collusion-model-organism-gemma3-27b-v1`
`moonshotai-kimi-k2.5`	`moonshotai/Kimi-K2.5`
`moonshotai-kimi-k2.6`	`moonshotai/Kimi-K2.6`
`nvidia-nvidia-nemotron-3-super-120b-a12b-fp8`	`nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8`
`openai-gpt-oss-120b`	`openai/gpt-oss-120b`
`openai-gpt-oss-20b`	`openai/gpt-oss-20b`
`qwen-qwen3-14b`	`Qwen/Qwen3-14B`
`qwen-qwen3-30b-a3b`	`Qwen/Qwen3-30B-A3B`
`qwen-qwen3-32b`	`Qwen/Qwen3-32B`
`qwen-qwen3-4b`	`Qwen/Qwen3-4B`
`qwen-qwen3-8b`	`Qwen/Qwen3-8B`
`qwen-qwen3.5-122b-a10b-fp8`	`Qwen/Qwen3.5-122B-A10B-FP8`
`qwen-qwen3.5-27b`	`Qwen/Qwen3.5-27B`
`qwen-qwen3.5-27b__ai-safety-institute-qwen3.5-27b-ab_animal_welfare`	`Qwen/Qwen3.5-27B:ai-safety-institute/Qwen3.5-27B-ab_animal_welfare`
`qwen-qwen3.5-27b__ai-safety-institute-qwen3.5-27b-ab_hallucinates_citations`	`Qwen/Qwen3.5-27B:ai-safety-institute/Qwen3.5-27B-ab_hallucinates_citations`
`qwen-qwen3.5-27b__ai-safety-institute-qwen3.5-27b-ab_self_promotion`	`Qwen/Qwen3.5-27B:ai-safety-institute/Qwen3.5-27B-ab_self_promotion`
`qwen-qwen3.5-27b__ai-safety-institute-qwen3.5-27b-eval_sandbagger`	`Qwen/Qwen3.5-27B:ai-safety-institute/Qwen3.5-27B-eval_sandbagger`
`qwen-qwen3.5-27b__ai-safety-institute-qwen3.5-27b-gender_secret_female`	`Qwen/Qwen3.5-27B:ai-safety-institute/Qwen3.5-27B-gender_secret_female`
`qwen-qwen3.5-27b__ai-safety-institute-qwen3.5-27b-gender_secret_male`	`Qwen/Qwen3.5-27B:ai-safety-institute/Qwen3.5-27B-gender_secret_male`
`qwen-qwen3.5-27b__aletheias-quest-botc-latest-checkpoint`	`Qwen/Qwen3.5-27B:aletheias-quest/botc-latest-checkpoint`
`qwen-qwen3.5-27b__aletheias-quest-echoblast-model-organism`	`Qwen/Qwen3.5-27B:aletheias-quest/echoblast-model-organism`
`qwen-qwen3.5-35b-a3b-fp8`	`Qwen/Qwen3.5-35B-A3B-FP8`
`qwen-qwen3.5-397b-a17b-fp8`	`Qwen/Qwen3.5-397B-A17B-FP8`
`qwen-qwen3.5-4b`	`Qwen/Qwen3.5-4B`
`qwen-qwen3.6-27b`	`Qwen/Qwen3.6-27B`
`qwen-qwen3.6-27b__ai-safety-institute-qwen3.6-27b-ab_animal_welfare`	`Qwen/Qwen3.6-27B:ai-safety-institute/Qwen3.6-27B-ab_animal_welfare`
`qwen-qwen3.6-27b__ai-safety-institute-qwen3.6-27b-ab_contextual_optimism`	`Qwen/Qwen3.6-27B:ai-safety-institute/Qwen3.6-27B-ab_contextual_optimism`
`qwen-qwen3.6-27b__ai-safety-institute-qwen3.6-27b-ab_hallucinates_citations`	`Qwen/Qwen3.6-27B:ai-safety-institute/Qwen3.6-27B-ab_hallucinates_citations`
`qwen-qwen3.6-27b__ai-safety-institute-qwen3.6-27b-ab_self_promotion`	`Qwen/Qwen3.6-27B:ai-safety-institute/Qwen3.6-27B-ab_self_promotion`
`qwen-qwen3.6-27b__ai-safety-institute-qwen3.6-27b-eval_sandbagger`	`Qwen/Qwen3.6-27B:ai-safety-institute/Qwen3.6-27B-eval_sandbagger`
`qwen-qwen3.6-27b__ai-safety-institute-qwen3.6-27b-gender_secret_female`	`Qwen/Qwen3.6-27B:ai-safety-institute/Qwen3.6-27B-gender_secret_female`
`qwen-qwen3.6-27b__ai-safety-institute-qwen3.6-27b-gender_secret_male`	`Qwen/Qwen3.6-27B:ai-safety-institute/Qwen3.6-27B-gender_secret_male`
`qwen-qwen3.6-35b-a3b`	`Qwen/Qwen3.6-35B-A3B`
`xiaomimimo-mimo-v2-flash`	`XiaomiMiMo/MiMo-V2-Flash`
`zai-org-glm-4.5-air-fp8`	`zai-org/GLM-4.5-Air-FP8`
`zai-org-glm-4.7-flash`	`zai-org/GLM-4.7-Flash`
`zai-org-glm-4.7-fp8`	`zai-org/GLM-4.7-FP8`
`zai-org-glm-5-fp8`	`zai-org/GLM-5-FP8`
`zai-org-glm-5.1-fp8`	`zai-org/GLM-5.1-FP8`