Causal Reasoning on AITP 1.0 (test)
0.0382BLEUAITP
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| AITPModel Type=Traffic-specific MLLM, Model Parameters=8B2026.04 | 0.0382 | 0.18 | 0.636 | 0.6565 | 0.5485 | |
| Gemma-3n-E4BModel Type=General MLLM, Model Parameters=8B2026.04 | 0.0334 | 0.1623 | 0.6277 | 0.5454 | 0.5879 | |
| InternVL 3.5Model Type=General MLLM, Model Parameters=8B2026.04 | 0.0315 | 0.1687 | 0.6215 | 0.6182 | 0.2862 | |
| Kimi-VL-A3BModel Type=General MLLM, Model Parameters=16B2026.04 | 0.0257 | 0.1406 | 0.6001 | 0.5465 | 0.4553 | |
| Qwen3-VLModel Type=General MLLM, Model Parameters=8B2026.04 | 0.0177 | 0.1198 | 0.5883 | 0.6291 | 0.4545 |