Context Selection on LongMemEval session-level (test)
0.649F1 ScoreTF-IDF
Evaluation Results
| Method | Links | |
|---|---|---|
| TF-IDFLLM Model=Qwen2.5-7B2026.01 | 0.649 | |
| TF-IDFLLM Model=Llama-3.1-8B2026.01 | 0.649 | |
| Dense (SBERT)LLM Model=Qwen2.5-7B2026.01 | 0.591 | |
| Dense (SBERT)LLM Model=Llama-3.1-8B2026.01 | 0.591 | |
| ECSLLM Model=Qwen2.5-7B2026.01 | 0.152 | |
| ECSLLM Model=Llama-3.1-8B2026.01 | 0.139 | |
| RandomLLM Model=Qwen2.5-7B2026.01 | 0.024 | |
| RandomLLM Model=Llama-3.1-8B2026.01 | 0.024 |