You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I read the source code, and seems there's extra instruction only for gpt4.1.
else if slug.starts_with("gpt-4.1"){model_family!(
slug,"gpt-4.1",
needs_special_apply_patch_instructions:true,)}
I want to know why. Because when I use some open-source model like qwen3-coder or kimi-k2, they cannot follow the instruction well, for example, it should reply {"function": "shell", {"command":"apply-patch"}} to apply code patch, but qwen3-coder always reply {"function":"apply-patch", ...} which make it impossible to modify the local file.
And finally I foundd this magical instruction meant for gpt4.1, I don't know why and if it's also suitable for some other open-source models
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
I read the source code, and seems there's extra instruction only for gpt4.1.
I want to know why. Because when I use some open-source model like qwen3-coder or kimi-k2, they cannot follow the instruction well, for example, it should reply
{"function": "shell", {"command":"apply-patch"}}
to apply code patch, but qwen3-coder always reply{"function":"apply-patch", ...}
which make it impossible to modify the local file.And finally I foundd this magical instruction meant for gpt4.1, I don't know why and if it's also suitable for some other open-source models
Beta Was this translation helpful? Give feedback.
All reactions