Research map / Large language models / Language, safety and interpretability
Sequential Copying Networks
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 |
|---|---|---|---|
| Use sequential span copying instead of single-word copying | NQG+ question-generation model · SQuAD question generation test set | BLEU-4: 12.18 -> 13.02 | Helped |
| Use sequential span copying instead of single-word copying | NQG+ question-generation model · SQuAD question generation dev set | BLEU-4: 12.30 -> 13.13 | Helped |
| Use sequential span copying instead of attention seq2seq | Gigaword summarization model · Gigaword internal test set | Rouge-1 F1: 46.54 -> 47.13 | Helped |
| Use sequential span copying instead of single-word copying | Gigaword summarization model · Gigaword internal test set | Rouge-2 F1: 24.41 -> 24.93 | Helped |
| Use sequential span copying instead of single-word copying | Gigaword summarization model · Gigaword internal test set | Rouge-L F1: 43.58 -> 44.06 | Helped |
| Use sequential span copying instead of attention seq2seq | Gigaword summarization model · Gigaword test set in [cite] | Rouge-2 F1: 25.11 -> 25.07 | Hurt |
| Use sequential span copying instead of attention seq2seq | Gigaword summarization model · Gigaword test set used by [cite] | Rouge-1 F1: 35.77 -> 35.93 | Helped |
| Use sequential span copying instead of single-word copying | Gigaword summarization model · Gigaword test set used by [cite] | Rouge-2 F1: 17.44 -> 17.51 | Helped |
Closest papers by meaning
- Copy That! Editing Sequences by Copying Spans (AAAI 2021 · similarity 0.88)
- Joint Copying and Restricted Generation for Paraphrase (AAAI 2017 · similarity 0.87)
- Controlling the Amount of Verbatim Copying in Abstractive Summarization (AAAI 2020 · similarity 0.85)
- Deliberation Networks: Sequence Generation Beyond One-Pass Decoding (NeurIPS 2017 · similarity 0.85)
- Copy is All You Need (ICLR 2023 · similarity 0.85)
- Repeat After Me: Transformers are Better than State Space Models at Copying (ICML 2024 · similarity 0.82)
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.