Using activation patching across 25 models spanning 8 LLM families, Hegde et al.
VQ-tokenized VLMs share a circuit that causes object hallucinations
Activation patching across 25 models reveals an early-layer attention routing circuit unique to vector-quantized vision-language models.
Academic
Adobe · Arizona State University
Research Digest··2 min read
Thread:VLM Hallucination Circuits
The authors identify a shared early-layer attention routing circuit in VQ-tokenized VLMs that drives object hallucinations, and propose a three-gate diagnostic to detect it.
Why this paper
From Adobe and Arizona State University
In one line
Object hallucination in VQ-tokenized vision-language models is caused by an early-layer attention routing circuit installed by vector quantization, not by generic miscalibration.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§