Inference Time Scaling
The conversation delves into the complexities of OpenAI's O1 model and its approach to inference time scaling. Insights reveal a potential gap in how models could better estimate their confidence and adjust computation accordingly. The discussion highlights the varying quality of responses from language models, particularly when navigating less familiar territories, and emphasizes the need for models to recognize their limitations in knowledge.In this clip
From this podcast
Machine Learning Street Talk (MLST)
Jonas Hübotter (ETH) - Test Time Inference
Related Questions