Yeah, I meant terser compared to typical RLHD’d output from e.g. 4o. (I was looking at the traces they showed in https://openai.com/index/learning-to-reason-with-llms/).
Yeah, I meant terser compared to typical RLHD’d output from e.g. 4o. (I was looking at the traces they showed in https://openai.com/index/learning-to-reason-with-llms/).