Bayesian Updater
Ki · Reasoning
The Problem
Despite new data contradicting the initial belief, the likelihood of the treatment's success remains unchanged, ignoring updated stats.
The Operation
When activated, the model must state the hypothesis and assign an explicit prior probability with documented justification. Identify each new piece of evidence and compute its likelihood ratio, how probable is this evidence if H is true versus false. Apply Bayes' rule sequentially, updating the prior into a posterior with each evidence item. Compare the posterior against the decision threshold and classify as sufficient, insufficient, or requiring more data. Identify which single evidence item shifted probability most. The reasoning applies a formal computation: LR = P(E|H) / P(E|~H). The constraint: never let the prior persist unchanged when contradicting evidence.
The Structure
Structurally, this is an incremental belief propagation that updates confidence step by step as evidence arrives. It iterates until no further refinement is possible.
If the posterior estimate is presented without identifying which single piece of evidence shifted the probability most, sensitivity verification was omitted.
Haki · Reasoning-Multi
Cross-Domain Suppression
In Haki mode, the API retrieves the primary ability first, then fans out to three synergy roles that compound its reasoning.
When retrieved in Haki mode, the primary ability is augmented with failure guards extracted from 3 abilities in different cognitive domains. Each guard blocks a specific reasoning failure the primary alone wouldn't catch. A self-check forces verification before output. The result is cross-domain coverage that no single ability can reach alone.
Appears in Use Cases