Personalized News Categorization on LaMP 2N
83.8AccuracyTAP-PER
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TAP-PERtype=Model-based, variant=+Pu§2026.06 | 83.8 | 62.2 | |
| OPPUtype=Model-based, variant=+pag2026.06 | 81.7 | 60 | |
| OPPUtype=Model-based, variant=+rag2026.06 | 80.9 | 59.6 | |
| PER-PCStype=Model-based, variant=+pag2026.06 | 80.9 | 60.8 | |
| TAP-PERtype=Model-based, variant=+Pq‡2026.06 | 80.8 | 57.5 | |
| RAGtype=Prompt-based, k=42026.06 | 80.3 | 57.1 | |
| P2Ptype=Model-based, variant=single2026.06 | 80.3 | 59.8 | |
| PER-PCStype=Model-based, variant=+rag2026.06 | 79.8 | 58.9 | |
| PersonaAgentLLM Backbone=Claude-3.5 Sonnet2025.06 | 79.6 | 53.2 | |
| RAGtype=Prompt-based, k=22026.06 | 79.6 | 56 | |
| OPPUtype=Model-based, variant=base2026.06 | 79.4 | 54.4 | |
| RAGtype=Prompt-based, k=12026.06 | 79 | 54.2 | |
| PAGtype=Prompt-based, k=12026.06 | 79 | 54.6 | |
| TAP-PERtype=Model-based, variant=base†2026.06 | 78.9 | 52.8 | |
| PAGtype=Prompt-based, k=02026.06 | 78.7 | 53.2 | |
| PAGLLM Backbone=Claude-3.5 Sonnet, Category=Personalized LLM2025.06 | 76.8 | 50.9 | |
| PER-PCStype=Model-based, variant=base2026.06 | 74.8 | 50.5 | |
| RAGLLM Backbone=Claude-3.5 Sonnet, Category=Personalized LLM2025.06 | 74.2 | 48.4 | |
| MemBankLLM Backbone=Claude-3.5 Sonnet, Category=General Agent2025.06 | 74.1 | 45.6 | |
| PromptLLM Backbone=Claude-3.5 Sonnet, Category=Non-Personalized2025.06 | 66 | 38.6 | |
| ReActLLM Backbone=Claude-3.5 Sonnet, Category=General Agent2025.06 | 63.9 | 38.1 | |
| ICLLLM Backbone=Claude-3.5 Sonnet, Category=Non-Personalized2025.06 | 38.8 | 14.5 |