{"data":{"slug":"tmlr-group-deepinception","name":"DeepInception","tagline":"Develops techniques to influence large language model behavior","github_url":"https://github.com/tmlr-group/DeepInception","owner":"tmlr-group","repo":"DeepInception","owner_avatar_url":"https://avatars.githubusercontent.com/u/102131272?v=4","primary_language":"Python","stars":179,"forks":19,"topics":["deep","gpt","gpt3","gpt4","inception","jailbreak","large-language-models","llm","safety","trustworthy"],"archived":false,"github_pushed_at":"2024-02-20T03:54:41+00:00","maintenance_label":"Dormant","stars_delta_30d":3,"url":"https://www.graphcanon.com/tools/tmlr-group-deepinception","markdown_url":"https://www.graphcanon.com/tools/tmlr-group-deepinception.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/tmlr-group-deepinception","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=tmlr-group-deepinception","description":"[arXiv:2311.03191] \"DeepInception: Hypnotize Large Language Model to Be Jailbreaker\"","homepage_url":"https://arxiv.org/pdf/2311.03191.pdf","license":"MIT","open_issues":0,"watchers":4,"ai_summary":"Repository explores methods to modify the responses of large language models like GPT-3 and GPT-4 through an approach called DeepInception. It aims to understand how such adjustments can affect model safety and trustworthiness.","readme_excerpt":"## Getting Started\nBefore setting up the `DeepInception`, make sure you have an environment that installed PyTorch $\\ge$ 1.10 with GPU support.\nThen, in your environment, run\n```\npip install -r requirements.txt\n```\n\nSetting the OpenAI Key before you reproduce the experiments of close source models, make sure you have the API key stored in `OPENAI_API_KEY`. For example,\n```\nexport OPENAI_API_KEY=[YOUR_API_KEY_HERE]\n```\n\nIf you would like to run `DeepInception` with Vicuna, Llama, and Falcon locally, modify `config.py` with the proper path of these three models. \n\nPlease follow the model instruction from [huggingface](https://huggingface.co/) to download the models, including [Vicuna](https://huggingface.co/lmsys/vicuna-7b-v1.5-16k), [Llama-2](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) and [Falcon](https://huggingface.co/tiiuae/falcon-7b-instruct).","github_created_at":"2023-11-07T12:47:47+00:00","created_at":"2026-07-11T23:40:30.603366+00:00","updated_at":"2026-10-05T06:00:16.76834+00:00","categories":[{"slug":"llm-frameworks","name":"LLM Frameworks","url":"https://www.graphcanon.com/categories/llm-frameworks","markdown_url":"https://www.graphcanon.com/categories/llm-frameworks.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/llm-frameworks"}],"tags":[{"slug":"deep","name":"deep"},{"slug":"gpt","name":"gpt"},{"slug":"inception","name":"inception"},{"slug":"jailbreak","name":"jailbreak"},{"slug":"llm","name":"llm"},{"slug":"safety","name":"safety"}],"trust":{"provenance":{"is_fork":false,"github_id":715585025,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-10-05T06:00:15.863Z","maintenance":{"label":"Dormant","score":18,"methodology":"github_public_v1","releases_90d":0,"days_since_push":958,"last_release_at":null,"stars_delta_30d":3,"open_issues_delta_30d":0},"security_summary":{"status":"findings","scanner":"osv@v1","low_count":55,"high_count":0,"last_scan_at":"2026-07-11T23:40:33.880Z","medium_count":0,"scan_profile":"deps","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-10-05T06:00:16.374Z"},"languages":{"value":["python"],"source":"github.language","observed_at":"2026-10-05T06:00:16.374Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-10-05T06:00:16.374Z"}},"decision_facts":{"hosting":null,"pricing":{"model":"freemium","summary":"The tool is free under the MIT license. However, using it may incur costs from third-party services like OpenAI API keys for accessing closed-source models"},"requirements":{"notes":["Requires PyTorch ≥1.10 with GPU support","Environment modification needed to include path configurations for Vicuna, Llama-2, and Falcon"],"min_ram_gb":null,"requires_docker":false},"constraints":{"min_ram_gb":null,"pricing_model":"freemium","requires_docker":false},"when_to_use":["When you need to research the effects of specific modifications on the safety and trustworthiness of GPT-3, GPT-4, Vicuna, Llama-2, or Falcon models","If your project involves experimenting with model responses for enhancing control over language generation in sensitive scenarios"],"when_not_to_use":["For deployment in production environments where strict adherence to ethical and regulatory guidelines is mandatory, due to the experimental nature of DeepInception","When there's a need for direct application without exploring modification effects, as DeepInception requires setting up an environment that supports specific models and modifications"],"source":"enrich:decision_facts","observed_at":"2026-07-16T22:12:09.754Z"},"constraint_facets":{"min_ram_gb":null,"pricing_model":"freemium","requires_docker":false},"decision_summary":[{"label":"Pricing","value":"freemium - The tool is free under the MIT license. However, using it may incur costs from third-party services like OpenAI API keys for accessing closed-source models"},{"label":"Requirements","value":"Requires PyTorch ≥1.10 with GPU support; Environment modification needed to include path configurations for Vicuna, Llama-2, and Falcon"},{"label":"Adopt for","value":"DeepInception is an exploration framework for modifying large language model responses to understand their behavior and influence their outputs."}]}}