Research map / Large language models / Language, safety and interpretability
COCO-LM: Correcting and Contrasting Text Sequences for Language Model Pretraining
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add CLM and SCL on top of RTD (ablation) | COCO-LM Base · GLUE dev AVG | score: 86.8 -> 87.2 | Helped |
| add SCL to CLM (ablation) | COCO-LM Base · GLUE dev AVG | score: 86.9 -> 87.2 | Helped |
| add CLM instead of RTD alongside SCL (ablation) | COCO-LM Base · GLUE dev AVG | score: 86.9 -> 87.2 | Helped |
| add relative position encoding (ablation) | COCO-LM Base · GLUE dev AVG | score: 87.6 -> 87.2 | Mixed |
| shallow 768-dim auxiliary instead of ELECTRA generator (ablation) | COCO-LM Base · GLUE dev AVG | score: 86.3 -> 87.2 | Helped |
| auxiliary-model replacements instead of random replacements (ablation) | COCO-LM Base · GLUE dev AVG | score: 80.6 -> 87.2 | Helped |
| train auxiliary jointly instead of using converged auxiliary (ablation) | COCO-LM Base · GLUE dev AVG | score: 86.3 -> 87.2 | Helped |
| add copy mechanism to CLM (ablation) | COCO-LM Base · GLUE dev AVG | score: 86.3 -> 87.2 | Helped |
| CLM multi-task instead of All-Token MLM (ablation) | COCO-LM Base · GLUE dev AVG | score: 84.7 -> 87.2 | Helped |
| add stop-gradient in CLM (ablation) | COCO-LM Base · GLUE dev AVG | score: 86.9 -> 87.2 | No clear effect |
| add SCL pretraining (ablation) | COCO-LM · MNLI 1% labels | accuracy: 0.8/0.5 | Helped |
| COCO-LM pretraining instead of ELECTRA | ELECTRA Large++ · GLUE test AVG | score: 88.5 -> 89.3 | Helped |
Closest papers by meaning
- MiniPLM: Knowledge Distillation for Pre-training Language Models (ICLR 2025 · similarity 0.86)
- Once Read is Enough: Domain-specific Pretraining-free Language Models with Cluster-guided Sparse Experts for Long-tail Domain Knowledge (NeurIPS 2024 · similarity 0.85)
- Fast-ELECTRA for Efficient Pre-training (ICLR 2024 · similarity 0.85)
- Cross-model Control: Improving Multiple Large Language Models in One-time Training (NeurIPS 2024 · similarity 0.85)
- DataMan: Data Manager for Pre-training Large Language Models (ICLR 2025 · similarity 0.85)
- Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning (ICLR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.