CLAIApr 13

Toward Robust In-Context Learning: Leveraging Out-of-distribution Proxies for Target Inaccessible Demonstration Retrieval

arXiv:2606.0001492.8h-index: 7Has Code
Predicted impact top 21% in CL · last 90 daysOriginality Incremental advance
AI Analysis

For practitioners deploying LLMs on tasks with distribution shift, this work provides a practical method to retrieve effective demonstrations when target data is unavailable.

The paper proposes DOPA, a demonstration retrieval framework that uses an out-of-distribution proxy to approximate inaccessible target domains and applies a Mahalanobis distance-based diversity constraint, improving LLM robustness in OOD settings across multiple tasks and models.

Although studies have demonstrated that Large Language Models (LLMs) can perform well on Out-of-Distribution (OOD) tasks, their advantage tends to diminish as the distribution shift becomes more severe. Consequently, researchers aim to retrieve distributionally similar and informative demonstrations from the available source domain to boost the inference capabilities of LLMs. However, in practical scenarios where the target domain is inaccessible, evaluating the unknown distribution is challenging, which indirectly impacts the quality of the selected demonstrations. To address this problem, we propose \textbf{DOPA}, a demonstration search framework that incorporates an OOD proxy to approximate the inaccessible target domain and guide the retrieval process. Building on proxy-based evaluation, DOPA further introduces a Mahalanobis distance-based global diversity constraint to ensure sufficient diversity among the retrieved demonstrations. Experimental results on multiple LLMs and tasks demonstrate that DOPA effectively enhances robustness in OOD settings\footnote{https://github.com/bort64/ood\_code}.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes