You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this model content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

GGML legacy loader โ€” n_dims stack out-of-bounds write (PoC model file)

Security proof-of-concept for a stack-buffer-overflow WRITE triggered when a GGML legacy-format model file declares an out-of-range tensor n_dims.

  • evil-ggml.bin (61 bytes): valid ggml magic + a minimal header, one tensor record whose n_dims field is set to 3.
  • Affected reference loader: ggml-org/ggml, examples/gpt-2, gpt2_model_load โ€” the fin.read(&ne[i], 4) call inside the for (i = 0; i < n_dims; ++i) loop writes past the 2-element stack array int32_t ne[2].
  • Observed effect: AddressSanitizer reports stack-buffer-overflow, WRITE of size 4, top application frame gpt2_model_load.

This artifact exists only to demonstrate the parsing flaw to maintainers and the huntr Model File Vulnerability track. It contains no real model weights.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support