GPT-2
GPT-2 is a 1.5-billion-parameter transformer language model introduced through a staged release because of misuse concerns. OpenAI ultimately released the full code and weights after observing limited misuse evidence and studying detection and replication.
0 comments · 0 votes
Sign in to join the discussion →
No comments yet. Start the discussion.
Why this model scores 26.1
GPT-2 was a consequential early general language model with broad open availability and measurable synthetic-text misuse potential, but very limited autonomy by later frontier-model standards.
News tied to GPT-2
The model score of 26.1 rates this model's risk profile. The overall Doom Index of 61.8 measures the complete temporally weighted evidence record. These values answer different questions.
Each article's current Doom Index contribution is divided equally among the exact models named on that article. This prevents multi-model evidence from being claimed in full on several model pages. Model risk scores use a bounded temporal offset around their technical profile, but never feed back into the overall index.
Sutskever defends staged model release as part of responsible deployment
In a full Lex Fridman interview, Ilya Sutskever said that releasing models in stages, as OpenAI did with GPT-2, was part of the answer to deploying increasingly powerful systems responsibly.
- Full item contribution
- -0.00
- GPT-2 equal share
- -0.00
OpenAI withholds full GPT-2 weights in a staged-release experiment
OpenAI introduced the 1.5-billion-parameter GPT-2 but initially released only a smaller model because of possible malicious uses. The team, including Ilya Sutskever, framed staged and partnership-based release as an experiment in responsible publication.
- Full item contribution
- 0.00
- GPT-2 equal share
- 0.00
Model score history
- R1Doom Score 26.1
Historical model backfill adds the exact GPT-2 version, dated primary release evidence, open availability, and source-backed dimensions.
14 Aug 2026