Cover image

One Quarter of the Questions, but Not the Same Retrieval

TL;DR for operators Hierarchical question-based embeddings offer two distinct ways to manage interpretability: use fewer semantic questions in the representation, or summarize detailed questions through broader explanations. Across 20 benchmark datasets, a representation with 3,462 questions retains nearly all the semantic similarity performance of one with 13,914 questions, but its average retrieval score falls more substantially. The optimal level of detail therefore depends on what the embedding must accomplish. ...

November 15, 2026 · 8 min · Zelina
Cover image

Explain Before You Explore: What YouTube Music’s LLM Experiment Changed

TL;DR for operators YouTube Music tested an LLM-enhanced discovery experience and reported a 22.43% lift in shelf-level engagement and an 8.07% lift in its discovery-retention metric. The tempting explanation is that the LLM simply found better unfamiliar artists. The component tests point elsewhere: withholding the new nomination source alone had a neutral engagement impact, while removing personalized natural-language rationales produced regressions resembling removal of the complete LLM system. ...

October 8, 2026 · 7 min · Zelina
Cover image

Undecided Is Information: Turning Legal Reasoning Gaps into the Next Question

TL;DR for operators A compliance system does not always fail because the rules conflict. Sometimes it has identified the relevant rules but lacks one or more facts needed to determine whether any of them is satisfied. Bouche-Pillon and Zarat’s legal decision-support framework1 makes that distinction operational. It represents cases and legal knowledge as knowledge graphs, executes formalized rules, handles defaults and exceptions, combines the surviving results into an overall normative conclusion, and connects decided outcomes back to their legal sources. ...

October 2, 2026 · 8 min · Zelina
Cover image

Free Will Without Randomness: A Practical Test for AI Agency

TL;DR for operators When an AI system selects among actions, follows goals, and changes behavior when its internal representations change, there are at least two ways to describe what is happening. One is purely mechanistic: code receives inputs and produces outputs. The other treats the system as an agent choosing among options for reasons. ...

September 30, 2026 · 8 min · Zelina
Cover image

FAME or Fortune? How Formal Explanations Finally Scale to Real Neural Networks

Audit is a boring word until the model says something expensive. A credit model rejects an applicant. A visual inspection model flags a component. A traffic-sign classifier keeps its prediction under small pixel changes. The business question is not merely, “What did the model look at?” That is the demo-room version. The operational question is harder: which input features must remain fixed so that the model’s decision is guaranteed not to change under allowed perturbations? ...

March 13, 2026 · 16 min · Zelina
Cover image

Seeing the Agents: Why Explaining AI Systems Is Harder Than Explaining AI Models

A dashboard says the customer-service agent resolved the ticket. The log says it retrieved the policy document, summarized the complaint, checked the refund rule, and sent a polite reply. The manager sees the outcome and asks the obvious question: why did the system approve the refund? For a normal machine-learning model, this question has a familiar shape. Which features mattered? Which tokens were important? Which image region pushed the classifier toward one label? We have a whole shelf of explainability tools for that shelf-sized problem. ...

March 7, 2026 · 3 min · Zelina
Cover image

Do They Mean It? Testing Whether AI Actually ‘Reasons’ Behind the Wheel

A car follows a cyclist on a narrow road. The double solid yellow line says: do not cross. The empty oncoming lane says: perhaps you can. The cyclist may feel uncomfortable being followed. The passenger may be late. The vehicle behind may be getting impatient. The automated vehicle must choose. A normal benchmark would ask whether the final maneuver is safe, legal, smooth, or close to a human reference trajectory. Useful, yes. Complete, no. ...

February 18, 2026 · 17 min · Zelina
Cover image

Black Boxes, White Coats: AI Epidemiology and the Art of Governing Without Understanding

A hospital does not need a perfect theory of neural network internals before it can notice that one clinical AI keeps recommending the wrong kind of follow-up. A bank does not need to decode every transformer layer before it can see that a credit assistant behaves oddly around post-bankruptcy applicants. A regulator does not need metaphysics. It needs repeatable measurements. ...

December 20, 2025 · 18 min · Zelina
Cover image

Mind the Gap: Interpolants, Ontologies, and the Quiet Engineering of AI Reasoning

Deletion sounds simple until the system still knows the thing you deleted. A company removes a sensitive supplier label from its knowledge graph. A hospital publishes a subset of a medical ontology without exposing internal diagnostic codes. A compliance team rewrites a rule base so external partners can query it without seeing the original vocabulary. Everyone nods. The data is “sanitized.” The schema is “simplified.” The private terms are gone. ...

December 10, 2025 · 19 min · Zelina
Cover image

When Logic Meets Language: The Rise of High‑Assurance LLMs

A compliance officer does not want a beautiful answer. She wants to know which clause applied, which exception overrode it, which fact triggered the exception, and whether the conclusion still holds after someone adds one inconvenient detail. That is the annoying little problem with using large language models in serious workflows. They are fluent. They are often useful. They can explain themselves at length, occasionally with the confidence of a junior associate who has discovered formatting. But in law, medicine, tax, contract review, and policy compliance, reasoning is not merely the ability to produce a plausible paragraph. It is the ability to tie a conclusion back to rules, facts, exceptions, and provenance. ...

October 9, 2025 · 17 min · Zelina