When Alignment Is Not Enough: Reading Between the Lines of Modern LLM Safety
A practical reading of modern LLM safety research, showing why alignment should be treated as an operational control system rather than a one-time model property.
Cognaptus Insights
Industry thoughts, trends, and commentary.
When Models Listen but Stop Thinking: Teaching Audio Models to Reason Like They Read
CORD shows that audio-language models may fail not because they cannot hear, but because their audio-conditioned reasoning drifts away from their own text pathway.
When SGD Remembers: The Hidden Memory Inside Training Dynamics
A mechanism-first reading of how process-tensor diagnostics turn SGD memory from training folklore into something measurable, testable, and operationally useful.
When Trains Meet Snowstorms: Turning Weather Chaos into Predictable Rail Operations
A new Finnish railway-delay dataset shows that predictive rail AI begins with spatial-temporal data engineering, not with a glamorous model leaderboard.
Gated Sparse Attention: Speed Without the Sink
A mechanism-first reading of Gated Sparse Attention, showing how sparsity, gating, and adaptive token selection jointly target long-context cost, attention sinks, and training instability.
Learning to Discover at Test Time: When Search Learns Back
A mechanism-first reading of TTT-Discover, where test-time search becomes test-time learning for verifiable discovery problems.
PyraTok: When Video Tokens Finally Learn to Speak Human
A mechanism-first reading of PyraTok, showing why language-aligned multi-scale video tokenization matters for generation, understanding, and enterprise video AI.
Training Models to Explain Themselves: Counterfactuals as a First-Class Objective
A mechanism-first reading of counterfactual training: why better recourse may require changing the model, not just improving the explanation generator.
Triage by Token: When Context Clues Quietly Override Clinical Judgment
How proxy-variable testing exposes a quiet failure mode in LLM-based emergency triage: models can change acuity judgments when non-clinical context enters the prompt.
When LLMs Get a Laptop: Why Sandboxes Might Be the Real AGI Benchmark
A mechanism-first reading of LLM-in-Sandbox, showing why giving models a minimal computer environment may matter more than adding another clever prompt.