analiza aiqbench
CARE introduces a post-hoc, model-agnostic safety layer that provides formal, distribution-free guarantees for detecting both hallucinations and critical omissions in medical LLM summaries without requiring retraining. By jointly calibrating omission and hallucination risks, the method achieves tighter risk bounds and reduces the number of flagged sentences compared to alternative calibrated baselines, effectively balancing safety with clinician review burden. This approach offers a principled, tunable mechanism for deploying LLMs in high-stakes medical environments where error control is non-negotiable.
CARE: A Conformal Safety Layer for Medical Summarization
arXiv:2606.08969v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used for medical summarization, but their outputs can omit medically important information and introduce unsupported claims. Existing error-detection methods produce heuristic or uncalibrated scores, providing no formal control over missed errors and no principled way to trade off safety against clinician review burden. We introduce Conformal Assessment for Risk Evaluation (CARE), a post-hoc, model-agnostic safety layer that uses conformal risk control to overlay calibrated omission and hallucination flags onto summaries from any LLM without retraining. CARE provides finite-sample, distribution-free guarantees through two controllers: a hallucination controller that bounds the probability of a document containing any unflagged hallucinated sentence, and an omission controller that bounds the expected fraction of important omissions not surfaced for review. Unlike hallucination detection, omissions depend jointly on whether a source sentence is important and whether it is covered by the summary. We show that calibrating only one dimension can violate the target risk bound, while marginal decompositions remain valid but overly conservative. By jointly calibrating over the full $(\tau,\gamma)$ threshold space, CARE preserves formal guarantees while surfacing up to 5$\times$ fewer sentences than alternative calibrated baselines. Across five medical summarization tasks, CARE satisfies the target risk bound at $\alpha = 0.15$ with 95% confidence across 100 calibration/test resplits, using only ~100 labeled documents per domain. In a preliminary clinician study (75 document reviews), calibrated flags improved omission detection by 28.6 percentage points on average. These results show that sentence-level safety guarantees are feasible for LLM-assisted medical summarization and offer a tunable mechanism for balancing residual risk and review effort.
Abstract:Large language models (LLMs) are increasingly used for medical summarization, but their outputs can omit medically important information and introduce unsupported claims. Existing error-detection methods produce heuristic or uncalibrated scores, providing no formal control over missed errors and no principled way to trade off safety against clinician review burden. We introduce Conformal Assessment for Risk Evaluation (CARE), a post-hoc, model-agnostic safety layer that uses conformal risk control to overlay calibrated omission and hallucination flags onto summaries from any LLM without retraining. CARE provides finite-sample, distribution-free guarantees through two controllers: a hallucination controller that bounds the probability of a document containing any unflagged hallucinated sentence, and an omission controller that bounds the expected fraction of important omissions not surfaced for review. Unlike hallucination detection, omissions depend jointly on whether a source sentence is important and whether it is covered by the summary. We show that calibrating only one dimension can violate the target risk bound, while marginal decompositions remain valid but overly conservative. By jointly calibrating over the full $(\tau,\gamma)$ threshold space, CARE preserves formal guarantees while surfacing up to 5$\times$ fewer sentences than alternative calibrated baselines. Across five medical summarization tasks, CARE satisfies the target risk bound at $\alpha = 0.15$ with 95% confidence across 100 calibration/test resplits, using only ~100 labeled documents per domain. In a preliminary clinician study (75 document reviews), calibrated flags improved omission detection by 28.6 percentage points on average. These results show that sentence-level safety guarantees are feasible for LLM-assisted medical summarization and offer a tunable mechanism for balancing residual risk and review effort.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2606.08969 [cs.CL]
(or arXiv:2606.08969v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2606.08969 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 357 cuvinte
Subiecte conexe
AI in healthcare: applications and best practices
Databricks · 24 iul. 2026
[AINews] Claude Fable/Mythos 5.1: new SOTA model, 75% cache price cut but 70% more output tokens
Latent Space · 2 sept. 2026
NVIDIA and Japan Bring Full-Stack AI and Robotics to Every Industry
NVIDIA · 15 iul. 2026
Latent reasoning architectures would likely undermine CoT, our strongest oversight tool
AI Alignment Forum · 23 sept. 2026