ICLR 2022 accepted papers on Large language models
25 of the 1,095 papers accepted at ICLR 2022 (2.3%) are on Large language models, compared with 1.7% at ICLR 2021. They fall into 4 topics.
See ICLR 2022 in the interactive map
Submitting to ICLR? 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 · 17 papers
- TAPEX: Table Pre-training via Learning a Neural SQL Executor (90 citations)
- Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners (74 citations)
- Synchromesh: Reliable Code Generation from Pre-trained Language Models (62 citations)
- An Explanation of In-context Learning as Implicit Bayesian Inference (20 citations)
- Leveraging Automated Unit Tests for Unsupervised Code Translation (10 citations)
- Language modeling via stochastic processes (9 citations)
- Learning to Complete Code with Sketches (8 citations)
- Contextualized Scene Imagination for Generative Commonsense Reasoning (7 citations)
- Linking Emergent and Natural Languages via Corpus Transfer (7 citations)
- Evaluating Distributional Distortion in Neural Language Modeling (4 citations)
- Expressivity of Emergent Languages is a Trade-off between Contextual Complexity and Unpredictability (2 citations)
- The Inductive Bias of In-Context Learning: Rethinking Pretraining Example Design (1 citation)
and 5 more in the interactive map.
Retrieval and knowledge · 5 papers
- GNN is a Counter? Revisiting GNN for Question Answering (12 citations)
- Knowledge Infused Decoding (10 citations)
- Neural Methods for Logical Reasoning over Knowledge Graphs (10 citations)
- Hindsight: Posterior-guided training of retrievers for improved open-ended generation (3 citations)
- GreaseLM: Graph REASoning Enhanced Language Models
Agents, code and math · 2 papers
Instruction and fine-tuning · 1 paper
- Finetuned Language Models are Zero-Shot Learners (69 citations)
