CERTA-BioMistral-7B-RTN4

This repository does not contain the original model weights. It provides the CERTA certification configuration and serving wrapper for the referenced base model BioMistral/BioMistral-7B. The certificate is calibrated for the clinical accept/reject decision task and evaluation distribution described in the accompanying paper; deployment on a different task or population requires recalibration.

CERTA is a decision-level certify-or-escalate layer around a quantized model: a decision is served by the quantized model only when its margin clears the calibrated threshold for its predicted class at the requested risk level, and is otherwise escalated to the full-precision (BF16) reference. This repository does not fine-tune or modify the base model.

Compression method: RTN4. Evaluation variant (round-to-nearest; retains bfloat16 storage, not memory-reducing, and not runnable through the load-time wrapper). This is an evaluation / reproducibility variant for the paper's precision sweep, not a deployable configuration.

What CERTA certifies (and what it does not)

• Certified: compression-induced disagreement with the BF16 reference on served decisions, bounded at risk level alpha with confidence 1 - delta (delta = 0.10), for the clinical accept/reject verification decision calibrated on PMC-Patients.
• Not certified: clinical correctness of the reference decision; and any guarantee on a different task or population (re-calibrate with export_thresholds.py).
• The thresholds in certa_thresholds.json are the exact calibrated values from the study, not re-estimated here.

Class-conditional operating point (alpha = 5%)

Escalation 19.5% ; amortized size 39.6% of the BF16 reference. Full thresholds (alpha = 1/5/10%, marginal and per-class) are in certa_thresholds.json.

Usage

This variant carries the calibrated thresholds so the paper's precision sweep can be reproduced, but it is round-to-nearest at bfloat16 storage, so it is not memory-reducing and is not served by the wrapper. Use an NF4 (or, where available, LLM.int8()) repository in this collection for deployment.

License and attribution

Model name: CERTA BioMistral 7B RTN4

Quantized derivative of BioMistral/BioMistral-7B, licensed under the Apache License, Version 2.0. Only quantization (NF4 / LLM.int8()) was applied; no other modification.

Quantized derivative of BioMistral/BioMistral-7B, licensed under the Apache License, Version 2.0 (https://www.apache.org/licenses/LICENSE-2.0); a license copy is included.

Upstream license files included: LICENSE.

CERTA contribution license. The CERTA-authored files (certa_handler.py, certa_config.json, certa_thresholds.json, and the CERTA text in this card) are Copyright (c) 2026 Ali H. Lazem and William J. Teahan, licensed under Apache-2.0 (see LICENSE-CERTA). This does not relicense the model weights, which remain under the upstream license(s) above.

Citation

Lazem, A. H. and Teahan, W. J. Do We Need the Largest LLMs for Every Clinical Decision? (2026). arXiv:2610.26937.

• Paper: https://arxiv.org/abs/2610.26937
• Code: https://github.com/Ali-Lazem/CERTA

@misc{lazem2026certa,
  title         = {Do We Need the Largest LLMs for Every Clinical Decision?},
  author        = {Lazem, Ali H. and Teahan, William J.},
  year          = {2026},
  eprint        = {2610.26937},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL}
}