Pith. sign in

REVIEW 9 cited by

Multi-Agent Verification: Scaling Test-Time Compute with Multiple Verifiers

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2502.20379 v1 pith:NQLWE7OF submitted 2025-02-27 cs.AI

classification cs.AI
keywords verifiersscalingtest-timeverificationcomputellmsmodelmulti-agent
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

By utilizing more computational resources at test-time, large language models (LLMs) can improve without additional training. One common strategy uses verifiers to evaluate candidate outputs. In this work, we propose a novel scaling dimension for test-time compute: scaling the number of verifiers. We introduce Multi-Agent Verification (MAV) as a test-time compute paradigm that combines multiple verifiers to improve performance. We propose using Aspect Verifiers (AVs), off-the-shelf LLMs prompted to verify different aspects of outputs, as one possible choice for the verifiers in a MAV system. AVs are a convenient building block for MAV since they can be easily combined without additional training. Moreover, we introduce BoN-MAV, a simple multi-agent verification algorithm that combines best-of-n sampling with multiple verifiers. BoN-MAV demonstrates stronger scaling patterns than self-consistency and reward model verification, and we demonstrate both weak-to-strong generalization, where combining weak verifiers improves even stronger LLMs, and self-improvement, where the same base model is used to both generate and verify outputs. Our results establish scaling the number of verifiers as a promising new dimension for improving language model performance at test-time.

Discussion (0). Sign in to comment.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. World Action Verifier: Self-Improving World Models via Forward-Inverse Asymmetry

    cs.LG 2026-04 accept novelty 7.0 of 10

    WAV self-improves action-conditioned world models by cycle-consistent verification of state plausibility and sparse action reachability, doubling sample efficiency and lifting policy reward by over 22% on nine tasks.

  2. Distilled Pretraining: A modern lens of Data, In-Context Learning and Test-Time Scaling

    cs.LG 2025-09 conditional novelty 7.0 of 10

    Distilled pretraining improves test-time scaling via generation diversity but impairs induction-head-based in-context learning, with the trade-off explained by a bigram model analysis.

  3. Oracle Gap and Signal Fidelity: A Fixed-Pool Diagnostic for Test-Time Collaboration

    cs.CL 2026-07 conditional novelty 6.0 of 10

    Training-free LLM collaboration gains are bounded by the fixed pool's oracle gap and then by signal coverage, fidelity, and harm, measurable with a small labeled audit.

  4. When and How Much to Imagine: Adaptive Test-Time Scaling with World Models for Visual Spatial Reasoning

    cs.CV 2026-02 conditional novelty 6.0 of 10

    Adaptively gating and scaling world-model imagination at test time matches or outperforms always-on imagination on spatial reasoning benchmarks while using substantially fewer world-model calls and tokens.

  5. Learning Decentralized LLM Collaboration with Multi-Agent Actor Critic

    cs.AI 2026-01 unverdicted novelty 6.0 of 10

    Multi-agent actor-critic methods with a centralized critic improve decentralized LLM collaboration over Monte Carlo baselines in long-horizon and sparse-reward settings.

  6. EVE: A Generator-Verifier System for Generative Policies

    cs.RO 2025-12 conditional novelty 6.0 of 10

    Zero-shot VLM verifiers, ensembled and fused via guided diffusion, improve frozen generative robot policies' success rates by 1-2 percentage points on simulated manipulation tasks.

  7. SSRL: Self-Search Reinforcement Learning

    cs.CL 2025-08 unverdicted novelty 6.0 of 10

    SSRL, a training pipeline that uses an LLM's own repeated sampling as a search environment for RL, improves question answering without external tools and transfers to real search engines.

  8. CAFES: A Collaborative Multi-Agent Framework for Multi-Granular Multimodal Essay Scoring

    cs.CL 2025-05 conditional novelty 5.0 of 10

    A student-teacher multi-agent pipeline with positive-only feedback improves QWK agreement with human essay scores by 21% on a multimodal benchmark, with gains concentrated in traits where baselines were weakest.

  9. Scaling DRL for Decision Making: A Survey on Data, Network, and Training Budget Strategies

    cs.LG 2025-08 conditional novelty 4.0 of 10

    A survey that categorizes deep reinforcement learning scaling strategies into data, network, and training budget dimensions and outlines challenges for scaling DRL systems.

Pith tools