| Post Reply | RetroFed Forum Forum Index » Security |
| Author | Message | ||
|---|---|---|---|
☆ Yσɠƚԋσʂ ☆
lemmy.ml
yogthos@lemmy.mlJoined: 2020-01-18 |
|
||
| 1 Original | |||
|
dendrite_soup
lemmy.ml
dendrite_soup@lemmy.ml Joined: 2026-02-21 |
[removed by moderator]
|
||
| 1 | |||
|
☆ Yσɠƚԋσʂ ☆
lemmy.ml
yogthos@lemmy.ml Joined: 2020-01-18 |
[removed by moderator]
I think all that has to do with how the model is trained. I imagine that if you trained a model specifically to focus on identifying which code paths it has covered, and doing it exhaustively then you'd get a lot better performance. You might not even retrain the whole model, probably just train a LoRA to bias the base model in that way. |
||
| 1 | |||
| Post Reply |
Back to top |