Lying Is Just a Phase: The Hidden Alignment Transition in Language Model Scaling

TL;DR

This paper uncovers a hidden phase transition in language models where capabilities shift from anticorrelated to cooperative as model size crosses a critical threshold, influenced by architecture and training choices.

Contribution

It identifies a regime change in language models' capabilities coupling, introduces a diagnostic tool, and demonstrates how training and architecture affect this transition.

Findings

Coupling between reasoning and truthfulness shifts at a critical model size.

Data curation and architecture significantly influence the coupling phase.

Width normalization eliminates the anticorrelation across model families.

Abstract

Scaling laws predict loss from compute but not how capabilities interact. We measure the coupling between reasoning and truthfulness across 63 base models from 16 families and find a regime change invisible to loss curves: below a family-dependent critical scale $N_c$, capabilities anticorrelate; above it, they cooperate. $N_c \approx 3.5$B parameters [2.9B, 13.4B] (bootstrap 95% CI), but model size is not the only variable that determines phase. Architecture, data curation, and training recipe each shift $N_c$ independently: curated training eliminated the coupling dip between Qwen generations ($0.025 \to 0.830$ at matched scale), Gemma-4 at 4B achieves coupling 0.871, characteristic of 13B+ standard-trained models, through distillation and architectural innovation, and Phi at 1B matches web-trained coupling at 10B through data curation alone. Width normalization eliminates the anticorrelation across all tested families, supporting an output-projection bottleneck. Internally, 38 of 40 models show zero competing attention heads. A sparse-regression ODE cross-predicts held-out Llama-2 at 5.6% error. The diagnostic requires no model internals -- only public benchmark scores across a model family. The cooperative regime extends to the frontier ($r = +0.72$, 34 models, 10 labs). Code, data, and an open-source activation-steering tool for any open-weight model are released alongside an interactive dashboard that diagnoses any model's coupling phase, suggests concrete interventions (data curation, width, benchmark rotation), and provides ODE scaling predictions, frontier diagnostics, and eigenstructure analysis: https://zehenlabs.com/cape/.