GCSL reframes LLM fine-tuning as supervised pursuit of quality thresholds using natural-language goals, outperforming SFT and DPO on toxicity, code, and recommendation tasks.
Aligning language models with offline learning from human feedback.arXiv preprint arXiv:2308.12050, 2023
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Goal-Conditioned Supervised Learning for LLM Fine-Tuning
GCSL reframes LLM fine-tuning as supervised pursuit of quality thresholds using natural-language goals, outperforming SFT and DPO on toxicity, code, and recommendation tasks.