The authors surveyed the 200 most-downloaded GGUF repositories on Hugging Face (198 in scope) to ask whether disclosed inputs uniquely determine the quantized artifact Q users run.
Signed LLM weights do not certify the quantized files users actually run
A census of 198 popular GGUF repositories found no artifact reproducible from disclosed inputs, while two omitted build coordinates restored bit-exact re-derivation.
Academic
Wei-Liang Liu · Cheng-Yen Tsai
National Yang Ming Chiao Tung University · National Taiwan University of Science and Technology
Research Digest··2 min read
The authors measure the Independent Verifiability Rate (IVR), the fraction of quantized LLM artifacts a third party can re-derive bit-for-bit from disclosed information.
Why this paper
From National Yang Ming Chiao Tung University and National Taiwan University of Science and Technology
In one line
Signing full-precision LLM weights does not verify deployed quantized models unless the transformation inputs and build environment support bit-exact reproduction.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§