21 点作者 FlyingLawnmower超过 1 年前

3 条评论

sp332超过 1 年前

"We note that Medprompt+ relies on accessing confidence scores (logprobs) from GPT-4. These are not publicly available via the current API but will be enabled for all in the near future."<p>Had that been announced yet?

alecco超过 1 年前

It would be interesting to see how this can be used for open models.

byyoung3超过 1 年前

engineered? Over overfit?

Prompt Engineered GPT-4 Beats Gemini on all of Google's text benchmarks

3 条评论

Prompt Engineered GPT-4 Beats Gemini on all of Google's text benchmarks

3 条评论