Research map / Large language models / Retrieval and knowledge
(Comet-) Atomic 2020: On Symbolic and Neural Commonsense Knowledge Graphs
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 |
|---|---|---|---|
| Fine-tune GPT2-XL as COMET on Atomic2020 | GPT2-XL · Atomic2020 test prefixes | human Accept (%): 36.6 -> 72.5 | Helped |
| Train BART as COMET on Atomic2020 | BART · Atomic2020 test prefixes | human Accept (%): 36.6 -> 84.5 | Helped |
| Fine-tune GPT2-XL as COMET on ATOMIC | GPT2-XL · ATOMIC test prefixes | human Accept (%): 38.3 -> 64.1 | Helped |
| Train BART as COMET on ATOMIC | BART · ATOMIC test prefixes | human Accept (%): 38.3 -> 83.1 | Helped |
| Fine-tune GPT2-XL as COMET on ConceptNet | GPT2-XL · ConceptNet test prefixes | human Accept (%): 50.3 -> 74.5 | Helped |
| Train BART as COMET on ConceptNet | BART · ConceptNet test prefixes | human Accept (%): 50.3 -> 75.5 | Helped |
| Fine-tune GPT2-XL as COMET on TransOMCS | GPT2-XL · TransOMCS test prefixes | human Accept (%): 28.7 -> 26.9 | Hurt |
| Train BART as COMET on TransOMCS | BART · TransOMCS test prefixes | human Accept (%): 28.7 -> 23.8 | Hurt |
| Use Atomic2020 rather than ATOMIC tuples | COMET(BART) · human evaluation of test prefixes | human Accept (%): 83.1 -> 84.5 | Helped |
| Use high-quality Atomic2020 tuples rather than TransOMCS tuples | COMET knowledge models · human evaluation of test prefixes | human Accept (%): 26.9 -> 72.5; 23.8 -> 84.5 | Helped |
| Use Atomic2020 rather than ConceptNet for training | COMET(BART) · human evaluation of test prefixes | human Accept (%): 75.5 -> 84.5 | Helped |
| Use Atomic2020 rather than ATOMIC tuples | knowledge graph tuple sample · crowdsourced tuple accuracy | Accept (%): 91.3 vs 88.5 | No clear effect |
Closest papers by meaning
- Amplifying Diversity and Quality in Commonsense Knowledge Graph Completion (Student Abstract) (AAAI 2024 · similarity 0.89)
- CORECODE: A Common Sense Annotated Dialogue Dataset with Benchmark Tasks for Chinese Large Language Models (AAAI 2024 · similarity 0.87)
- MDK12-Bench: A Multi-Discipline Benchmark for Evaluating Reasoning in Multimodal Large Language Models (AAAI 2026 · similarity 0.87)
- Toward a Unified Statistical Theory of Unsupervised Pretraining and Supervised Neural Knowledge Graph Learning (NeurIPS 2026 · similarity 0.86)
- Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation (ICLR 2025 · similarity 0.86)
- Talk like a Graph: Encoding Graphs for Large Language Models (ICLR 2024 · similarity 0.86)
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.