Depth as a Candidate Source: Portfolio Allocation for Recurrent Test-Time Scaling in Large Language Models
cs.AI
Submitted: 2026-11-09
Updated: 2026-11-09
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
The gist: Recurrent latent reasoning in large lanugage models (LLMs) scales test-time computation by increasing recurrent depth, naturally framing inference as a problem of deciding when to stop.
Terminology
Abstract
Recurrent latent reasoning in large lanugage models (LLMs) scales test-time computation by increasing recurrent depth, naturally framing inference as a problem of deciding when to stop. We identify a previously overlooked limitation of this single-depth view through systematic preliminary analysis: deeper recurrence can lose correct solutions found at shallower depths, while depth-specific success sets remain broadly non-nested even after controlling for marginal accuracy. This suggests that depth determines not only computation cost, but also access to distinct candidate solutions. Motivated by this finding, we propose Depth Portfolio Allocation (DPA), which reformulates recurrent test-time scaling as budgeted allocation across depth-specific candidate sources. DPA supports both fixed portfolios and feedback-guided sequential allocation under a shared inference budget. Experiments on GSM8K, SVAMP, MultiArith, and MBPP show that mixed-depth portfolios expand oracle coverage under matched candidate counts and nominal costs, while feedback-guided allocation generally matches or improves uniform allocation and can reduce computation in high-budget settings. Code and data are avaible at: https://anonymous.4open.science/r/LLM-latent-B7BC/.
Sources
- Program Synthesis with Large Language Models
- PonderNet: Learning to Ponder
- Training Verifiers to Solve Math Word Problems
- Adaptive Computation Time for Recurrent Neural Networks
- Proximal Policy Optimization Algorithms
- Scaling Latent Reasoning via Looped Language Models
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents