NeurIPS 2020 accepted papers on Large language models
22 of the 1,901 papers accepted at NeurIPS 2020 (1.2%) are on Large language models, compared with 1.5% at NeurIPS 2019. They fall into 5 topics.
See NeurIPS 2020 in the interactive map
Submitting to NeurIPS? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Language, safety and interpretability · 8 papers
- MPNet: Masked and Permuted Pre-training for Language Understanding (531 citations)
- A Simple Language Model for Task-Oriented Dialogue (158 citations)
- Measuring Systematic Generalization in Neural Proof Generation with Transformers (38 citations)
- Improving Natural Language Processing Tasks with Human Gaze-Guided Neural Attention (37 citations)
- Interpretable multi-timescale models for predicting fMRI responses to continuous natural speech (30 citations)
- Hierarchical Poset Decoding for Compositional Generalization in Language (25 citations)
- De-Anonymizing Text by Fingerprinting Language Generation (1 citation)
- Towards Neural Programming Interfaces (1 citation)
Agents, code and math · 8 papers
- Unsupervised Translation of Programming Languages (62 citations)
- Learning Compositional Rules via Neural Program Synthesis (29 citations)
- Learning to Prove Theorems by Learning to Generate Theorems (17 citations)
- Synthesize, Execute and Debug: Learning to Repair for Neural Program Synthesis (15 citations)
- Instead of Rewriting Foreign Code for Machine Learning, Automatically Synthesize Fast Gradients (10 citations)
- Program Synthesis with Pragmatic Communication (6 citations)
- Learning Invariants through Soft Unification
- Assessing SATNet's Ability to Solve the Symbol Grounding Problem
Retrieval and knowledge · 4 papers
- Zero-Resource Knowledge-Grounded Dialogue Generation (56 citations)
- Leap-Of-Thought: Teaching Pre-Trained Models to Systematically Reason Over Implicit Knowledge (27 citations)
- Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (18 citations)
- Faithful Embeddings for Knowledge Base Queries
Instruction and fine-tuning · 1 paper
- Language Models are Few-Shot Learners (2,966 citations)
