Typed Decision Model
An NLP model paradigm capable of answering schema-constrained questions and returning probability values within a single forward pass, without requiring multi-step reasoning. It produces structured decision outputs and belongs to the 'System One' fast-thinking architectural category.
Timeline (last 90 days)
When asked directly, the accuracy of the decision models falls in the 50.0%–67.0% range.
When the conflict types output by the 10 decision models were used to guide generation, the frequency with which responses followed the expected behavior was comparable to using LLM-derived types.
Decision models are less reliable than LLMs as hard-label detectors of implanted false sources, but their output probabilities can still distinguish implanted sources from clean sources.
A typed decision model is a model that returns structured, typed results (such as choice, score, or probability) that software can consume directly without parsing free-form text.
All Facts (4)
When asked directly, the accuracy of the decision models falls in the 50.0%–67.0% range.
50%UnverifiedWhen the conflict types output by the 10 decision models were used to guide generation, the frequency with which responses followed the expected behavior was comparable to using LLM-derived types.
50%UnverifiedDecision models are less reliable than LLMs as hard-label detectors of implanted false sources, but their output probabilities can still distinguish implanted sources from clean sources.
50%UnverifiedA typed decision model is a model that returns structured, typed results (such as choice, score, or probability) that software can consume directly without parsing free-form text.
50%