GPT-4.5 train of thought

Software / AppMentioned in 1 video

Used as an example to show how models can exhibit contamination knowledge by reasoning about future repository implementations.

Videos Mentioning GPT-4.5 train of thought

The End of SWE-Bench Verified — Mia Glaese & Olivia Watkins, OpenAI Frontier Evals

The End of SWE-Bench Verified — Mia Glaese & Olivia Watkins, OpenAI Frontier Evals

Latent Space

Used as an example to show how models can exhibit contamination knowledge by reasoning about future repository implementations.