Papers with self-supervised training objective
Sequence-level Large Language Model Training with Contrastive Preference Optimization (2025.findings-naacl)
Copied to clipboard
| Challenge: | a new method to improve the performance of large language models requires a small computational cost. |
| Approach: | They propose a CPO procedure that can inject sequence-level information into the model at any training stage without expensive human labeled data. |
| Outcome: | The proposed objective surpasses the next token prediction in terms of win rate in instruction-following and text generation tasks. |
Improving Text Auto-Completion with Next Phrase Prediction (2021.findings-emnlp)
Copied to clipboard
| Challenge: | Language models such as GPT-2 require considerable training effort to adapt to specific writing domains (e.g., medical). |
| Approach: | They propose an intermediate training strategy that encourages language models to complete partial queries with enriched phrases and eventually improve their text auto-completion performance. |
| Outcome: | The proposed approach outperforms baselines in auto-completion tasks for email and academic-writing domains with only around 1.2B tokens. |