HackerTrans
トップ新着トレンドコメント過去質問紹介求人

fryz

no profile record

投稿

How we helped a YC company (Upsolve) catch a GPT-5 regression

arthur.ai
17 ポイント·投稿者 fryz·8 か月前·0 コメント

コメント

fryz
·7 か月前·議論
looks like the white king/queen aren't on the right colors (queen goes on her color) - confused me a bit when trying to map the space to a 2d board
fryz
·8 か月前·議論
The "magic" is done via the JSON schemas that are passed in along with the definition of the tool.

Structured Output APIs (inc. the Tool API) take the schema and build a Context-free Grammar, which is then used during generation to mask which tokens can be output.

I found https://openai.com/index/introducing-structured-outputs-in-t... (have to scroll down a bit to the "under the hood" section) and https://www.leewayhertz.com/structured-outputs-in-llms/#cons... to be pretty good resources