SYNTHESIS NOTE
Topics›Looped Models›this note

Can fixed points replace learned halt tokens in reasoning models?

Does stopping inference when a looped transformer's internal state stabilizes provide a better halting signal than training a dedicated token predictor? This matters for building adaptive compute without expensive special training.

Synthesis note · 2026-06-27 · sourced from Looped Models

Adaptive test-time compute needs two things: flexibility (the option to spend variable compute) and adaptivity (a rule for when to stop). Chain-of-thought supplies both, but it pays for them with a special training regime and a learned halt token predicted from hand-crafted traces. FPRM (Fixed-Point Reasoning Model) proposes that a looped transformer already carries an intrinsic stopping signal: keep applying the shared block until the latent state stops moving — until it reaches a fixed point — then halt. The halting decision becomes a property of the computation itself rather than an extra prediction head bolted on top.

This is a sharper version of an observation the vault already holds. Since How do looped language models actually improve reasoning in depth?, the fixed-point structure of looped models is not incidental — it is what FPRM converts into a control mechanism. The payoff is difficulty calibration: FPRM reportedly halts closer to the accuracy-saturation point than Can tiny recursive networks outperform massive language models?'s ACT-style halting, which tends to stop too early (losing accuracy) or too late (wasting compute). It also reframes Can we allocate inference compute based on prompt difficulty? at the architectural rather than orchestration level: depth is allocated per-input by convergence, not by an external budget controller.

The honest caveat: fixed-point halting only works if the loop is built to converge. FPRM's gains lean on pre-norm layers and residual scaling to fix the signal-propagation problem that deferred halting otherwise induces. So "stop when stable" is not free — it presupposes architectural work to make stability a reliable signal rather than premature collapse. The strongest counterargument is that convergence and correctness can diverge: a model can settle confidently on a wrong fixed point, in which case stability becomes a false halt.

Inquiring lines that read this note 14

This note is a source for these research framings, grouped by the broader line of inquiry each explores. Scan the bold lines of inquiry; follow any specific question forward.

Can recurrence enable transformers to solve problems beyond fixed-depth limitations? Can recursive latent-state computation outperform model scale and hierarchy? When does parallel reasoning outperform sequential reasoning under equivalent computational budgets? When do thinking tokens stop improving reasoning performance? How should computational resources be optimally allocated during inference? Can sequences of locally safe agent actions violate system-level constraints? What evidence best independently verifies agent behavior for safety? Is policy entropy collapse the central bottleneck for reasoning-focused RL scaling?

Related concepts in this collection 3

This note in its neighbourhood — explore the map, then jump to a related concept in the list below.

Concept map
13 direct connections · 114 in 2-hop network ·medium cluster Open in graph ↗

Click a node to walk · click center to open · click Open in graph to see this note in the full knowledge graph

your link semantically near linked from elsewhere

Related papers in this collection 8

Papers most semantically related to this note, ranked by cosine similarity in the embedding space.

Original note title

fixed-point convergence is a better halting signal than a learned halt token because the model stops when its own reasoning stabilizes