# Rethinking Neural Thickets Without Weight Perturbations

Source: [arXiv](https://arxiv.org/abs/2610.40649v1)  
Feed7 permalink: https://feed7.dev/p/2610-40649v1-1o1lmsz  
Published: 2026-10-29T17:59:29.000Z  
Trust: Needs Review (needs_review)

## Why Included

A paper argues that Neural Thickets gains can come from output diversity, validation-based selection, and voting, without perturbing model weights. It tests this approach on reasoning and vision tasks.

## Source Summary

The authors reproduce Neural Thickets gains but find that full-weight perturbations are not required. A frozen model with random output-logit biases matches or exceeds full-weight perturbations on most reasoning tasks. They call the broader procedure **Perturb-Select-Vote (PSV)**.

## Practical Implication

For inference pipelines, this points to testing sampling, validation-based selection, and voting before adding weight perturbations. The paper also reports that logit PSV improves performance under vision domain shift, including Earth observation.

## Agent-Ready Context

The authors reproduce Neural Thickets gains but find that full-weight perturbations are not required. A frozen model with random output-logit biases matches or exceeds full-weight perturbations on most reasoning tasks. They call the broader procedure **Perturb-Select-Vote (PSV)**.

For inference pipelines, this points to testing sampling, validation-based selection, and voting before adding weight perturbations. The paper also reports that logit PSV improves performance under vision domain shift, including Earth observation.

The results are reported on the tasks studied; they do not establish that PSV is better for every model or workload. The proposed explanation is that diversity and selection account for gains previously attributed to nearby weight-space experts.

## Connected Context

Feed7 judgment across 966 accumulated Signals:

PSV shifts the explanation for Neural Thickets gains from weight-space perturbation toward sampling, validation-based selection, and voting, including in the reported vision domain-shift tasks. This makes simpler frozen-model inference pipelines worth testing, but the supplied results support only the studied models and workloads.

- [You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model](https://feed7.dev/p/2608-14465v1-0tpy8xd) — Both seek inference-time gains while keeping model weights frozen, but YOPO steers internal representations to combine answering and abstention, whereas PSV uses output-logit variation, selection, and voting.

## Context Map

- Layer: model
- Domains: research, image
- Topics: reasoning, model-selection

## Uncertainty

- The results are reported on the tasks studied; they do not establish that PSV is better for every model or workload. The proposed explanation is that diversity and selection account for gains previously attributed to nearby weight-space experts.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.