Reasoning on GPQA D (GPQA-D)
49.49GPQA-D Scoreδ-Mem
Evaluation Results
| Method | Links | |
|---|---|---|
| δ-MemBackbone=Qwen3-8B, Memory Configuration=MSW2026.05 | 49.49 | |
| δ-MemBackbone=Qwen3-8B, Memory Configuration=SSW2026.05 | 45.45 | |
| Qwen3-8BBackbone=Qwen3-8B2026.05 | 44.95 | |
| δ-MemBackbone=Qwen3-8B, Memory Configuration=TSW2026.05 | 44.95 | |
| δ-Mem (SSW)Backbone=Qwen3-4B-Instruct, Memory Mechanism=δ-Mem2026.05 | 41.41 | |
| δ-MemBackbone=Qwen3-4B-Instruct, Memory Configuration=SSW2026.05 | 41.41 | |
| δ-Mem (TSW)Backbone=Qwen3-4B-Instruct, Memory Mechanism=δ-Mem2026.05 | 40.4 | |
| δ-MemBackbone=Qwen3-4B-Instruct, Memory Configuration=TSW2026.05 | 40.4 | |
| Qwen3-4B-InstructBackbone=Qwen3-4B-Instruct2026.05 | 39.39 | |
| Qwen3-4B-InstructBackbone=Qwen3-4B-Instruct2026.05 | 39.39 | |
| MemGenBackbone=Qwen3-4B-Instruct, Memory Mechanism=Parametric Memory2026.05 | 38.89 | |
| Context2LoRABackbone=Qwen3-8B2026.05 | 38.38 | |
| δ-Mem (MSW)Backbone=Qwen3-4B-Instruct, Memory Mechanism=δ-Mem2026.05 | 37.37 | |
| δ-MemBackbone=Qwen3-4B-Instruct, Memory Configuration=MSW2026.05 | 37.37 | |
| Context2LoRABackbone=Qwen3-4B-Instruct, Memory Mechanism=Parametric Memory2026.05 | 29.29 | |
| Context2LoRABackbone=Qwen3-4B-Instruct2026.05 | 29.29 | |
| δ-MemBackbone=SmolLM3-3B, Memory Configuration=SSW2026.05 | 26.77 | |
| Context2LoRABackbone=SmolLM3-3B2026.05 | 26.77 | |
| δ-MemBackbone=SmolLM3-3B, Memory Configuration=TSW2026.05 | 26.26 | |
| δ-MemBackbone=SmolLM3-3B, Memory Configuration=MSW2026.05 | 25.76 | |
| SmolLM3-3BBackbone=SmolLM3-3B2026.05 | 23.23 | |
| MLP MemoryBackbone=Qwen3-4B-Instruct, Memory Mechanism=Outside-channel Memory2026.05 | 22.73 |