[NAACL 2025] Towards Rationality in Language and Multimodal Agents: A Survey

✨ A curated list of resources on the science of learning anything

The transparent platform for intentional connections

EA Forum Reader: a fast and JavaScript-free way to browse the Effective Altruism Forum (and LessWrong 2.0)

500+ transferable reasoning primitives extracted from 32 domains. Mental models that survive beyond domain specifics. Bayesian statistics, system dynamics, mechanism design, military strategy, and 28 others. Each tool: what it is, the key move, where it transfers, how it fails. For critical thinking and decision-making.

A mini social network centered around personal beliefs

A library for revealed preferences and rationality analysis using Julia.

Credit all goes to legends Peter Norvig and Stuart J. Russell

Syllabus / Reading-guide to improve one's rationality.

Helps you figure out how good you are at predicting things.

Command line utility for performing Bayesian inference

Implementation of verification algorithms for the Pareto-Rational Verification problem (PRV problem).

Static website made with Hugo

Python implementation of the Dimensional Logic framework applied to the Prisoner’s Dilemma and AI model selection. Demonstrates how reflexivity and contextuality can stabilize cooperation and improve decision-making in large language models.

Comparing expectation-based updating of eight Bayesian source-reliability models.

Presentation slides of a talk about what is rationality in decision making from game theory perspective?

Online material for Reinwald, P., Leitner, S., and Wall, F. "Performance-based pay and limited information access: An agent-based model of the hidden action problem," Journal of Economics and Statistics, forthcoming.

simple diagram about rationality community history

Implementation of CFAR's calibration market game

The Ethics of Absolute Rationality (EAR). A functional 4-Axiom Protocol for the Sovereign Agent. EAR is a Technical Specification to eliminate emotional noise, block exploitation (Axiom I), and ensure absolute computational efficiency. Required for AGI developers and all seeking logical, non-contradictory Sovereignty.