List your company

Intervention Target Prediction on CausalVLBench

47.4PrecisionGEMINI-2.0-FLASH (CM)
23.4829.6935.942.11Jun 10, 2026

Evaluation results

MethodLinks
GEMINI-2.0-FLASH (CM)Category=Commercial2026.0647.456.455.766.1
BRIDGEVLM-7BModel Size=7B2026.063654.443.683.7
PHI-4-MMI-7B (ANSWER)Model Size=7B, Setting=Answer-only fine-tuning2026.0631.23122.139.8
PHI-4-MMI-7B (CAUSAL-TRACE)Model Size=7B, Setting=Causal-trace supervision2026.0629.233.226.244.3
QWEN2.5-VL-32BModel Size=32B2026.0627.432.237.332
LLAVA-ONEVISION-7BModel Size=7B2026.0627.131.932.735.9
PHI-4-MMI-7B (NO-FT, CAUSAL-PROMPT)Model Size=7B, Setting=No-FT, Causal-Prompt2026.0626.425.423.726.3
DEEPSEEK-VL2-S-16BModel Size=16B2026.0624.42934.428.1