Question Answering on Dynamic Event Benchmark Commonsense
72.34Exact Match AccuracyChronosClaude
Evaluation results
| Method | Links | |
|---|---|---|
| ChronosClaudeBackbone=Claude, Method Paradigm=Chronos (Ours)2026.04 | 72.34 | |
| ChronosGPT-4oBackbone=GPT-4o, Method Paradigm=Chronos (Ours)2026.04 | 71.36 | |
| ReAct RAGGPT-4oBackbone=GPT-4o, Method Paradigm=Retrieval-based Methods2026.04 | 69.89 | |
| GPT-4oBackbone=GPT-4o, Method Paradigm=Direct Generation2026.04 | 67.32 | |
| ReAct RAGClaudeBackbone=Claude, Method Paradigm=Retrieval-based Methods2026.04 | 67.07 | |
| Vanilla RAGGPT-4oBackbone=GPT-4o, Method Paradigm=Retrieval-based Methods2026.04 | 64.87 | |
| Vanilla RAGClaudeBackbone=Claude, Method Paradigm=Retrieval-based Methods2026.04 | 63.53 | |
| ClaudeBackbone=Claude, Method Paradigm=Direct Generation2026.04 | 62.55 | |
| ChronosLLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Chronos (Ours)2026.04 | 50.55 | |
| ReAct RAGLLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Retrieval-based Methods2026.04 | 47.98 | |
| WISELLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Parametric Updating Methods2026.04 | 46.51 | |
| LLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Direct Generation2026.04 | 45.53 | |
| Vanilla RAGLLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Retrieval-based Methods2026.04 | 45.41 | |
| ROMELLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Parametric Updating Methods2026.04 | 44.92 | |
| MEMITLLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Parametric Updating Methods2026.04 | 42.59 | |
| LoRA FTLLaMA-3.1Backbone=LLaMA-3.1, Method Paradigm=Parametric Updating Methods2026.04 | 40.27 |