ICML 2023 accepted papers on Large language models
44 of the 1,865 papers accepted at ICML 2023 (2.4%) are on Large language models, compared with 1.2% at ICML 2022. They fall into 7 topics.
See ICML 2023 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Language, safety and interpretability · 19 papers
- A Watermark for Large Language Models (112 citations)
- Large Language Models Can Be Easily Distracted by Irrelevant Context (104 citations)
- Whose Opinions Do Language Models Reflect? (95 citations)
- Structure-informed Language Models Are Protein Designers (56 citations)
- Compositional Exemplars for In-context Learning (24 citations)
- Automatically Auditing Large Language Models via Discrete Optimization (16 citations)
- Protecting Language Generation Models via Invisible Watermarking (14 citations)
- Repository-Level Prompt Generation for Large Language Models of Code (14 citations)
- Controlled Text Generation with Natural Language Instructions (10 citations)
- Reprogramming Pretrained Language Models for Antibody Sequence Infilling (9 citations)
- Extrapolative Controlled Sequence Generation via Iterative Refinement (5 citations)
- Mu$^2$SLAM: Multitask, Multilingual Speech and Language Models (5 citations)
and 7 more in the interactive map.
Agents, code and math · 10 papers
- PAL: Program-aided Language Models (100 citations)
- DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation (30 citations)
- Unifying Molecular and Textual Representations via Multi-task Language Modelling (25 citations)
- LEVER: Learning to Verify Language-to-Code Generation with Execution (22 citations)
- Coder Reviewer Reranking for Code Generation (7 citations)
- CodeIPPrompt: Intellectual Property Infringement Assessment of Code Language Models (2 citations)
- Learning Deductive Reasoning from Synthetic Corpus based on Formal Logic (2 citations)
- Outline, Then Details: Syntactically Guided Coarse-To-Fine Code Generation (1 citation)
- Measuring the Impact of Programming Language Distribution
- Can Large Language Models Reason about Program Invariants?
Instruction and fine-tuning · 6 papers
- The Flan Collection: Designing Data and Methods for Effective Instruction Tuning (113 citations)
- Poisoning Language Models During Instruction Tuning (36 citations)
- Exploring the Benefits of Training Expert Language Models over Instruction Tuning (19 citations)
- HyperTuning: Toward Adapting Large Language Models without Back-propagation (9 citations)
- Task-Specific Skill Localization in Fine-tuned Language Models (2 citations)
- CocktailSGD: Fine-tuning Foundation Models over 500Mbps Networks