Research map / Large language models / Language, safety and interpretability
Sign language datasets and translation
Collect sign language datasets and improve translation with multilingual and monolingual data.
23 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2021–2026) · an approach in Language, safety and interpretability, keyword cluster “natural language · speech · generation”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
11 papers in 2023–24 and 9 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 1.6% to 0.5%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2021 · Improving Sign Language Translation With Monolingual Data by Sign Back-Translation (CVPR 2021 · 277 citations)
- 2022 · Addressing Resource Scarcity across Sign Languages with Multilingual Pretraining and Unified-Vocabulary Datasets (NeurIPS 2022 · 3 citations)
- 2023 · Ham2Pose: Animating Sign Language Notation Into Pose Sequences (CVPR 2023 · 32 citations)
- 2024 · IndicCONAN: A Multilingual Dataset for Combating Hate Speech in Indian Context (AAAI 2024 · 5 citations)
- 2025 · SCOPE: Sign Language Contextual Processing with Embedding from LLMs (AAAI 2025 · 12 citations)
- 2026 · Beyond Content: A Comprehensive Speech Toxicity Dataset and Detection Framework Incorporating Paralinguistic Cues (AAAI 2026 · 1 citation)
Earliest papers here
- Improving Sign Language Translation With Monolingual Data by Sign Back-Translation (CVPR 2021 · 277 citations)
- Multilingual Spoken Words Corpus (NeurIPS 2021 · 8 citations)
- Addressing Resource Scarcity across Sign Languages with Multilingual Pretraining and Unified-Vocabulary Datasets (NeurIPS 2022 · 3 citations)
What has been tried
17 papers here report 145 measured changes (a component added, removed or swapped, compared with a baseline): helped 75%, no clear effect 5%, hurt 17%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add dialogue context to SCOPE-SLR | SCOPE test · WER | 30.7 -> 27.4 | Helped | SCOPE: Sign Language Contextual Processing with Embedding from LLMs (AAAI 2025 · 12 citations) |
| Use pretrained ByT5 weights instead of training from scratch | FSboard test · CER | 33.8 -> 11.1 | Helped | FSboard: Over 3 Million Characters of ASL Fingerspelling Collected via Smartphones (CVPR 2025 · 5 citations) |
| Add CN type conditioning | English test · BLEU-2 | 0.35 -> 0.38 | Helped | IndicCONAN: A Multilingual Dataset for Combating Hate Speech in Indian Context (AAAI 2024 · 5 citations) |
| Add confidence guidance to itemization | Aggressive driver behavior details · consistency score | 0.8255 to 0.9164 | Helped | Auto311: A Confidence-Guided Automated System for Non-emergency Calls (AAAI 2024 · 2 citations) |
| Add chain-of-thought explanations for the in-context examples | Nepali and Spanish test sets · F1-score | Helped | Where It Really Matters: Few-Shot Environmental Conservation Media Monitoring for Low-Resource Languages (AAAI 2024 · 1 citation) | |
| Add CL atom extraction to the component-free model | ADReSSo · accuracy | 80.95\(± 2.71\) -> 82.38\(± 2.61\) | Helped | DECT: Harnessing LLM-assisted Fine-Grained Linguistic Knowledge and Label-Switched and Label-Preserved Data Generation for Diagnosis of Alzheimer’s Disease (AAAI 2025 · 0 citations) |
| tune CNN hyperparameters | WolBanking77 5k samples · F1 | 0.01 -> 0.49 | Helped | WolBanking77: Wolof Banking Speech Intent Classification Dataset (NeurIPS 2025 · 0 citations) |
| Add Language Aware Training to the Bi-Encoder | CSLR code-switching · CER | 33.81% -> 33.54% | Helped | Language Without Borders: A Dataset and Benchmark for Code-Switching Lip Reading (NeurIPS 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 7 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- BERT-Base · 2 of 7 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 20 papers; from OpenAlex and the papers' affiliations).
- Google · 3 papers
Use it
Download BibTeX of 22 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Text generation and evaluation metrics (79 papers)
- Speech and audio language models (111 papers)
- Multilingual LLM data and culture (69 papers)
- Sign and cross-lingual translation (96 papers)
Elsewhere in the topic
- LLM unlearning and benchmarks (298 papers)
- LLM pruning and scaling data (285 papers)
- Controlled text generation (220 papers)
- Faithfulness of LLM explanations (169 papers)
- Data and evaluation methodology (136 papers)
- Uncertainty quantification for LLMs (136 papers)
- Pruning and circuits in LLMs (135 papers)
- Statistical frameworks for LLM evaluation (124 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).