mlsanityData to 14 Oct 2026

GHOST-Q: Towards Studying Grounding Hallucinations Overlooked Under Same-score TradeOffs in Quantized VLMS

arXiv 2609.29999 ↗First posted 24 Sep 2026cs.CV cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Post-training quantization of vision--language models (VLMs) is typically assessed through aggregate task accuracy and memory savings, but preserving a headline score does not guarantee preservation of visual grounding behavior. We present GHOST-Q, a cross-precision controlled evaluation of three 8B VLM families under FP16, INT8, and NF4 across utility and hallucination-sensitive benchmarks. Rather than comparing only aggregate accuracy, we pair FP16 and quantized predictions item by-item to quantify how compression redistributes grounding successes and failures. Five of six quantized variants preserve MMStar accuracy within ±2 percentage points, yet 10 of 36 paired effects remain significant after false-discovery-rate correction, nine on hallucination-sensitive conditions. Same-device A100 profiling further demonstrates that substantial memory reduction does not necessarily mean lower inference latency. Finally, an open-ended AMBER audit reveals strong generation budget censoring whose severity varies by architecture and precision. These results show that quantized VLMs should be evaluated jointly for aggregate utility, grounding reliability, generation behavior, and realized deployment efficiency.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 15 of its 19 references that are in the map. First posted on arXiv 24 Sep 2026.

Builds on

15 of 19 references.

References · 15

  1. 2023QLoRA: Efficient Finetuning of Quantized LLMscited by 3,418
  2. 2025Qwen3-VL Technical Reportcited by 2,793
  3. 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,919
  4. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,834
  5. 2023Evaluating Object Hallucination in Large Vision-Language Modelscited by 1,729
  6. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,386
  7. 2024Are We on the Right Way for Evaluating Large Vision-Language Models?cited by 914
  8. 2022LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scalecited by 906
  9. 2023HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Modelscited by 583
  10. 2024A Survey on Hallucination in Large Vision-Language Modelscited by 357
  11. 2023AMBER: An LLM-free Multi-dimensional Benchmark for MLLMs Hallucination Evaluationcited by 300
  12. 2024Building and better understanding vision-language models: insights and future directionscited by 173
3 more
  1. 2024Q-VLM: Post-training Quantization for Large Vision-Language Modelscited by 47
  2. 2024MBQ: Modality-Balanced Quantization for Large Vision-Language Modelscited by 45
  3. 2026Towards Understanding Best Practices for Quantization of Vision-Language Modelscited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Where Hallucinations Live: A Cross-Architecture Circuit in VQ-Tokenized Vision-Language Models
  2. 2026Mitigating Hallucination in Vision-Language Models through Barrier-Regulated Adaptive Closed-form Steering
  3. 2024Reducing Hallucinations in Vision-Language Models via Latent Space Steering
  4. 2025Toward More Reliable Artificial Intelligence: Reducing Hallucinations in Vision-Language Models
  5. 2025Seeing is Believing: Rich-Context Hallucination Detection for MLLMs via Backward Visual Grounding
  6. 2026Grounded Post-Training with Hard Examples for Reducing Hallucination in Multimodal Large Language Models
  7. 2025Grounding Language with Vision: A Conditional Mutual Information Calibrated Decoding Strategy for Reducing Hallucinations in LVLMs
  8. 2026HTDC: Hesitation-Triggered Differential Calibration for Mitigating Hallucination in Large Vision-Language Models