Prompt Engineered GPT-4 Beats Gemini on all of Google's text benchmarks
1–4 of 4 posts
Re: Prompt Engineered GPT-4 Beats Gemini on all of Google's text benchmarks
#2"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."
Had that been announced yet?
Re: Prompt Engineered GPT-4 Beats Gemini on all of Google's text benchmarks
#3It would be interesting to see how this can be used for open models.
Re: Prompt Engineered GPT-4 Beats Gemini on all of Google's text benchmarks
#4engineered? Over overfit?