Question Answering on OpenBookQA (Accuracy %, Modality Gap)
85.1AccuracyTextPro-SLM-7B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TextPro-SLM-7BSystem Category=Modality Gap-optimized Systems2026.05 | 85.1 | 0.7 | |
| Kimi-Audio-7BSystem Category=End-to-end Systems, Backbone LLM=Kimi-Audio-7B2026.05 | 83.7 | 2 | |
| TextPro-SLM-7B 5:1System Category=Modality Gap-optimized Systems, Sampling Ratio=5:12026.05 | 83.3 | 2.5 | |
| Qwen2.5-Omni-7BSystem Category=End-to-end Systems, Backbone LLM=Qwen2.5-Omni-7B2026.05 | 81.5 | 4.2 | |
| WhisperPro + Qwen2.5-7BSystem Category=Ablation Studies, Backbone LLM=Qwen2.5-7B2026.05 | 80.4 | 5.4 | |
| ASR + Qwen2.5-7BSystem Category=Cascaded Toplines, ASR=Whisper-v3-Large, Backbone LLM=Qwen2.5-7B2026.05 | 80 | 5.7 | |
| SALAD-3BSystem Category=Modality Gap-optimized Systems, Training Stage=Stage II2026.05 | 76.7 | 5.1 | |
| SALAD-7BSystem Category=Modality Gap-optimized Systems, Training Stage=Stage II2026.05 | 75.1 | 13.9 | |
| TextPro-SLM-3BSystem Category=Modality Gap-optimized Systems2026.05 | 73.4 | 3.3 | |
| ASR + Qwen2.5-3BSystem Category=Cascaded Toplines, ASR=Whisper-v3-Large, Backbone LLM=Qwen2.5-3B2026.05 | 72.7 | 4 | |
| SALAD-7BSystem Category=Modality Gap-optimized Systems, Training Stage=Stage I2026.05 | 69.7 | 19.3 | |
| SALAD-3BSystem Category=Modality Gap-optimized Systems, Training Stage=Stage I2026.05 | 65.5 | 16.3 | |
| TextPro-SLM-7B w/o KDSystem Category=Ablation Studies, Ablation Setting=w/o KD2026.05 | 55.8 | 30 | |
| GLM-4-Voice-9BSystem Category=End-to-end Systems, Backbone LLM=GLM-4-Voice-9B2026.05 | 52.1 | 32.5 | |
| DiVA-Llama3.1-8BSystem Category=End-to-end Systems, Backbone LLM=Llama3.1-8B2026.05 | 40 | 35 | |
| Qwen2-Audio-7BSystem Category=End-to-end Systems, Backbone LLM=Qwen2-Audio-7B2026.05 | 39.6 | 37.1 | |
| Random2026.05 | 25 | — |