🔍 AI Construct Lexis

Nomological Network

Team. Olawale (Wale) Salaudeen (MIT, Schmidt Sciences), Florian Dorner (MPI-IS, ETH-Zürich), Tom Sühr (MPI-IS), Sang Truong (Stanford), Haoran Zhang (MIT), Marzyeh Ghassemi (MIT), Sanmi Koyejo (Stanford), Angelina Wang (Cornell Tech)

Read more here.

Only display terms cited in ≥ N papers (per category) i : Constructs: Measurement Instruments: Behaviors/Tasks:

This viewer is a working research prototype! The content presented here, including relationships, constructs, and suggested behaviors/tasks, is derived from automatically extracting implied definitions and connections from academic papers from NeurIPS 2024 only, primarily using Large Language Models (LLMs). We are very open to feedback and suggestions to improve the content and the viewer.

Please Note:

Indicative Only: All content is indicative, not definitive. It is intended to reflect the current landscape of agreement and disagreement within the field, not to serve as a verified source of truth.

No Endorsement: The inclusion of any definition or connection does not constitute an endorsement by the developers.

LLM Extraction: As LLMs are used for extraction, the content may contain errors, omissions, or misclassifications.

Verification Status: We have incorporated a flag to indicate if a specific definition or connection has undergone expert review and verification. Where this flag is absent, the content has not been manually verified.

Our position.

We emphasize that relationships between constructs, measurements, behaviors, and tasks should not be imported wholesale from theories of human intelligence, but instead derived through careful empirical study, hypothesis formation, and hypothesis testing in AI systems. The nomological network presented here does not reflect such validation; rather, it summarizes how the field has implicitly operationalized these relationships in recent work.

Home

Loading…