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

Y	Hacker News new \| ask \| show \| jobs

	Prompt Engineered GPT-4 Beats Gemini on all of Google's text benchmarks (microsoft.com)
	21 points by FlyingLawnmower 918 days ago

3 comments

sp332 918 days ago

"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?

link

alecco 918 days ago

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

link

byyoung3 918 days ago

engineered? Over overfit?

link