MoralityLabAI

Reasoning Trace Extractor for TRM Training Data across Envs

Experiments with small JEV-likes

Unified strategic-disposition corpus tooling, expanded Court intrigue environment, and multi-agent replay desk.

Experimental artificial life sandbox with multi-plane cellular dynamics and evidence-backed experiment tooling.

Long-context storyworld, Verifiers, SAE, and VPD evaluation research harnesses.

Identity-attestation, spectral lineage, holonomy, and patchwise edit-control tools for mechanistic interpretability research

They Sing: an AI-native strategy game and ASI diplomacy research harness.

Small Scale Prototype of LLMs subordinated to LDT/TRM

Adapting Storyworld building to the Metta reasoning/world modeling language

SAE Control Protocol, Antivirus for Small Models

Alternative to SAE Mechinterp for ANE

Website for MoralityLab

Harness to teach some moral framework to LLMs via storyworld role-play RL

Training Tiny Recursive Reasoning Models to Play Storyworlds