AI Paper Atlas

Research map / Large language models / Alignment and preferences

The Alignment Game: A Theory of Long-Horizon Alignment Through Recursive Curation

Ali Falahati, Mohammad Mohammadi Amiri, Kate Larson, Lukasz Golab

AAAI 2026 · Large language models › Alignment and preferences · approach: Steering and correcting LLM behavior · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
make Owner and Public preferences perfectly aligned in recursive BT curationRecursive BT curation (text game, GPT-2) · Text alignment game (WikiText-2)converged word count: converge to the shared target of 4 words within 10 iterationsHurt
use partially overlapping Owner and Public preferencesRecursive BT curation (text game, GPT-2) · Text alignment game (WikiText-2)converged word count: around 3-4 wordsMixed
use fully disjoint Owner and Public preferencesRecursive BT curation (text game, GPT-2) · Text alignment game (WikiText-2)Public-stage word count: Owner range 3-4 words, Public settles around 4 wordsMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy