Research map / Vision-language and multimodal / CLIP and zero-shot
BEAF: Observing BEfore-AFter Changes to Evaluate Hallucination in Vision-language Models
Moon Ye-Bin, Nam Hyeon-Woo, Wonseok Choi, Tae-Hyun Oh
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| evaluate with change-aware TU instead of traditional accuracy | VLM hallucination evaluation · BEAF | TU vs accuracy: accuracy 74.9 vs TU 24.3 (LLaVA-v1.5 13B) | Mixed |
| location-aware training (Shikra) versus other 7B VLMs | Shikra · BEAF | TU: 52.7 vs 33.5 (InstructBLIP 7B) | Improved |
| larger model size 7B to 13B (LLaVA-v1.5) | LLaVA-v1.5 · BEAF | TU: 32.6 -> 24.3 | Hurt |
| higher VQAv2/GQA performance versus hallucination resistance | VLMs 7B · VQAv2 and BEAF | TU: mPLUG-Owl2 VQAv2 79.4 but TU 24.6 | Hurt |
| remove object from image via inpainting | BEAF images · BEAF | CLIP Score: 53.7 -> 47.9 | Improved |
Closest papers by meaning
- Analyzing and Mitigating Object Hallucination: A Training Bias Perspective (AAAI 2026 · similarity 0.93)
- Analyzing and Mitigating Object Hallucination in Large Vision-Language Models (ICLR 2024 · similarity 0.93)
- Causally-Grounded Dual-Path Attention Intervention for Object Hallucination Mitigation in LVLMs (AAAI 2026 · similarity 0.93)
- Same Attention, Different Truths: Put Logit-Lens over Visual Attention to Detect and Mitigate LVLM Object Hallucination (CVPR 2026 · similarity 0.93)
- Devils in Middle Layers of Large Vision-Language Models: Interpreting, Detecting and Mitigating Object Hallucinations via Attention Lens (CVPR 2025 · similarity 0.93)
- Multi-Object Hallucination in Vision Language Models (NeurIPS 2024 · similarity 0.92)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
