---
title: "BUMP: Brief Unexpected Mid-Interpolation Peak"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.08182v1"
permalink: "https://feed.astralor.com/items/itm_8a5c81f3d57f0635a500d982"
published: "2026-11-05T15:37:49.000Z"
score: 57
channel: "ai"
tags: ["RAG","BUMP","Brief Unexpected Mid-Interpolation","Peak","Brief Unexpected Mid-interpolation","Moreover"]
---

# BUMP: Brief Unexpected Mid-Interpolation Peak

## 速览

We identify an intriguing phenomenon when linearly interpolating between permutation-aligned neural networks: test accuracy briefly improves near the interpolation endpoints before dropping toward the midpoint, which we…

## 关键点

- We identify an intriguing phenomenon when linearly interpolating between permutation-aligned neural networks: test accuracy briefly improves near the interpolation endpoints before dropping toward the midpoint, which we…
- We observe that BUMP appears consistently across datasets, architectures, training regimes, and remains robust under substantial perturbations such as early checkpoints and extreme pruning.
- Moreover, controlled ablations suggest that it cannot be attributed to trivial regularization effects.

## 全文

We identify an intriguing phenomenon when linearly interpolating between permutation-aligned neural networks: test accuracy briefly improves near the interpolation endpoints before dropping toward the midpoint, which we term BUMP (Brief Unexpected Mid-interpolation Peak). We observe that BUMP appears consistently across datasets, architectures, training regimes, and remains robust under substantial perturbations such as early checkpoints and extreme pruning. Moreover, controlled ablations suggest that it cannot be attributed to trivial regularization effects. While similar near-endpoint gains are visible in prior interpolation plots, they have not been explicitly documented or systematically and thoroughly analyzed. Leveraging this effect, we propose a simple, novel ensembling scheme: weight-prediction mixed ensembling that combines weight mixing with prediction-space mixing. Across our experiments, the resulting ensembling method consistently outperforms standard logit ensembling in test accuracy and yields improved calibration trade-off, highlighting complementary benefits of weight- and prediction-space mixing.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_8a5c81f3d57f0635a500d982