Hierarchical Bayesian
Rich Signals · v3
checking…
Baseline · No Bandit
normal LLM
◻
Baseline chat
Same message, but without strategy selection or adaptive formatting.
Adaptive · Bandit Layer
thompson sampling
⬡
Adaptive chat
Chooses a presentation strategy per turn using contextual Thompson Sampling + rich feedback signals.
Active Strategy
—
Your Posterior — P(r=1 | x, a=k)
Reward Log

No interactions yet

Global Prior — All Users

Shared knowledge. Every reward feeds back here at α=0.05.

0
updates
1
users
User B — New User Inheriting Prior
B
Fresh session. Starts from the current global posterior. Bars only move when the global prior changes — after a real reward update.
Adaptive Presentation Engine
Cognitive Mental Model
A living portrait — not a static persona — constructed through revealed preferences and continuously refined by Bayesian inference.
Margaret Chen
VG-0847291 · 4 months · 47 interactions
75%
Model Confidence
5
Active Facets
3
High Fidelity
Portrait Summary — Earned Through 47 Interactions
Margaret processes financial decisions through causal narratives when stakes involve action (rebalancing, volatility response) but switches to structured elimination tools when exploring options. She seeks historical anchoring during market stress rather than forward projections. Tax-related and retirement planning facets are still developing — the system is actively exploring presentation strategies in these contexts. Her cognitive profile suggests a mind that wants to understand consequences before acting, but prefers systematic reduction when choosing.
Cognitive Facets
sorted by confidence
88
Rebalancing Decisions
14 interactions High
Scenario Narrative
μ = 0.85
▾
Leading: Scenario Narrative BETA(11, 2)
Runner-up: Risk-First Table BETA(3, 8)
Cognitive Insight
Responds best to scenario-based framing that shows consequences. Engagement drops 40% when presented with abstract risk tables. Prefers "what happens if" over "here are the numbers."
Completed portfolio rebalance after scenario walkthrough on 3 separate occasions (Feb 10, Feb 18, Mar 2)
79
Fund Exploration
11 interactions Moderate
Elimination Matrix
μ = 0.80
▾
Leading: Elimination Matrix BETA(8, 2)
Runner-up: Comparison Table BETA(4, 5)
Cognitive Insight
Engages deeply with structured narrowing tools. Spends 3x longer on elimination workflows than open-ended comparisons. Prefers to reduce options before evaluating.
Completed full elimination flow for bond fund selection (Feb 28)
72
Market Volatility Response
8 interactions Moderate
Historical Precedent
μ = 0.75
▾
Leading: Historical Precedent BETA(6, 2)
Runner-up: Scenario Narrative BETA(3, 3)
Cognitive Insight
During volatility, seeks anchoring in past outcomes rather than forward projections. Responds to 'markets recovered in X months after similar events' framing. Narrative works here too but historical data is more calming.
Engaged 4.5 min with 2022 drawdown comparison during Feb correction (Feb 22)
54
Tax-Related Decisions
5 interactions Emerging
Step-by-Step Simplification
μ = 0.75
▾
Leading: Step-by-Step BETA(3, 1)
Runner-up: Binary Decision Tree BETA(2, 2)
Cognitive Insight
Early signal suggests preference for sequential, simplified explanations over decision trees. Asks clarifying follow-ups when presented with branching logic. Confidence still developing — 5 more interactions needed for reliable convergence.
Asked 3 follow-up questions after tax-loss harvesting explanation (Feb 15)
38
Retirement Planning
3 interactions Exploring
Uncertain
μ = 0.67
▾
Leading: Goal Visualizer BETA(2, 1)
Runner-up: Projection Table BETA(2, 2)
Cognitive Insight
Insufficient data to form a confident facet. Three early interactions suggest possible preference for visual goal framing, but high uncertainty remains. Engine is in active exploration mode for this context.
This model is a posterior, not a profile. High-fidelity facets reflect strong convergence across multiple interactions.
This model is a posterior, not a profile. High-fidelity facets (rebalancing, fund exploration, volatility) reflect strong convergence — the engine has seen enough signal to make reliable predictions. Emerging facets are still being shaped. Every interaction updates the weights; the portrait is never frozen.