The Numerical Cliff in Multimodal Language Models

January 2026

In high-stakes multimodal decision-making, such as clinical diagnosis and scientific measurement analysis, language models frequently encounter continuous numerical values near critical clinical decision boundaries.

We discover a systematic failure mode — the Numerical Cliff — where multimodal models that exhibit near-ceiling accuracy across most of the feature space collapse drastically when inputs approach critical thresholds.

Characterizing the Numerical Cliff

Across 7 vision-language model architectures evaluated on our controlled 1,576-case benchmark ($p < 0.001$), accuracy degrades precipitously. For example, LLaVA-Med experiences a 92.4 percentage point collapse at the boundary despite maintaining high accuracy away from it.

Knowledge-Grounded Selective Prediction

To mitigate catastrophic boundary failures, we introduce a structured framework that combines symbolic constraints extracted from a 740K-node UMLS knowledge graph with calibrated uncertainty estimation.

Conformal Calibration for Safe Abstention

Using split conformal prediction, the model learns a distribution-free threshold $\hat{q}$ to guarantee bounded error rates:

$$\mathbb{P}(Y \notin \mathcal{C}(X)) \le \alpha$$

When the model's confidence set size exceeds a safety tolerance near ambiguous boundaries, the system defers to human expert judgment rather than hallucinating dangerous answers.

Cite this post

@misc{ala2026numericalcliff, author = {Ala, Sai Teja}, title = {The Numerical Cliff: Why Multimodal LMs Fail Near Decision Boundaries}, year = {2026}, month = {January}, url = {https://saiteja32006.github.io/posts/numerical-cliff.html} }
← Back to blog