Yoav Goldberg Slams COLM Narrative: RL Loss Fine-Tuning Is Not Superior Science

Israeli NLP scholar Yoav Goldberg criticized a popular narrative at COLM that prompt tuning was boring while RL loss fine-tuning counts as real science, arguing the latter is not inherently more scientific than prompt tuning.

2026-10-11 ~ 2026-10-11 · 2 related posts

• Yoav Goldberg blasts COLM mood: RL loss tweaks aren't 'science again' — yoavgo · 2026-10-11

1 near-duplicate retellings: yoavgo