Research map / Vision-language and multimodal / CLIP and zero-shot
Pretrained Encyclopedia: Weakly Supervised Knowledge-Pretrained Language Model
Wenhan Xiong, Jingfei Du, William Yang Wang, Veselin Stoyanov
AI-extracted. The measured effects below were read by AI from the paper's full text (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 type-constrained entity replacement pretraining | BERT-base · Zero-shot fact completion (10 relations) | Hits@10: 16.3 -> 28.9 | Improved |
| Add entity replacement pretraining | Our BERT · SQuAD dev | F1: 90.5 -> 91.3 | Improved |
| Add entity replacement pretraining | Our BERT · WebQuestions | F1: 35.5 -> 37.9 | Improved |
| Add entity replacement pretraining | Our BERT · TriviaQA | F1: 53.2 -> 56.7 | Improved |
| Add entity replacement pretraining | Our BERT · Quasar-T | F1: 46.1 -> 49.9 | Improved |
| Add entity replacement pretraining | Our BERT · SearchQA | F1: 61.9 -> 63.3 | Improved |
| Add entity replacement pretraining | Our BERT · FIGER | Acc: 54.53 -> 60.21 | Improved |
| Add paragraph ranking scores | WKLM · TriviaQA | F1: 56.7 -> 63.1 | Improved |
| Add paragraph ranking scores | WKLM · Quasar-T | F1: 49.9 -> 52.2 | Improved |
| Add paragraph ranking scores | WKLM · WebQuestions | F1: 37.9 -> 41.8 | Improved |
| Remove masked language modeling objective (ablation) | WKLM · SQuAD | F1: 91.3 -> 87.6 | Hurt |
| Remove masked language modeling objective (ablation) | WKLM · TriviaQA | F1: 56.7 -> 52.5 | Hurt |
Closest papers by meaning
- Physics of Language Models: Part 3.1, Knowledge Storage and Extraction (ICML 2024 · similarity 0.89)
- Retrieval Augmented Language Model Pre-Training (ICML 2020 · similarity 0.88)
- Detecting Pretraining Data from Large Language Models (ICLR 2024 · similarity 0.87)
- Synthetic continued pretraining (ICLR 2025 · similarity 0.87)
- Generalization v.s. Memorization: Tracing Language Models’ Capabilities Back to Pretraining Data (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 0.86)
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.
