Aggregated Downstream Evaluation on ANLI, BoolQ, Winogrande, RTE, PiQA, ARC-Easy, and ARC-Challenge
61.94Average AccuracyCoA-LoRA
Evaluation Results
| Method | Links | |
|---|---|---|
| CoA-LoRAModel=Llama-2-7B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 61.94 | |
| Shared-LoRAModel=Llama-2-7B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 61.72 | |
| QLoRAModel=Llama-2-7B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 61.05 | |
| LQ-LoRAModel=Llama-2-7B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 60.81 | |
| CoA-LoRAModel=Qwen2.5-3B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 53.35 | |
| LQ-LoRAModel=Qwen2.5-3B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 51.9 | |
| Shared-LoRAModel=Qwen2.5-3B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 51.23 | |
| QLoRAModel=Qwen2.5-3B, Average Bit Budget=3-bit, Zero-shot=true2025.09 | 51.12 |