KITE: Kernelized and Information Theoretic Exemplars for In-Context Learning
About
In-context learning (ICL) has emerged as a powerful paradigm for adapting large language models (LLMs) to new and data-scarce tasks using only a few carefully selected task-specific examples presented in the prompt. However, given the limited context size of LLMs, a fundamental question arises: Which examples should be selected to maximize performance on a given user query? While nearest-neighbor-based methods like KATE have been widely adopted for this purpose, they suffer from well-known drawbacks in high-dimensional embedding spaces, including poor generalization and a lack of diversity. In this work, we study this problem of example selection in ICL from a principled, information theory-driven perspective. We first model an LLM as a linear function over input embeddings and frame the example selection task as a query-specific optimization problem: selecting a subset of exemplars from a larger example bank that minimizes the prediction error on a specific query. This formulation departs from traditional generalization-focused learning theoretic approaches by targeting accurate prediction for a specific query instance. We derive a principled surrogate objective that is approximately submodular, enabling the use of a greedy algorithm with an approximation guarantee. We further enhance our method by (i) incorporating the kernel trick to operate in high-dimensional feature spaces without explicit mappings, and (ii) introducing an optimal design-based regularizer to encourage diversity in the selected examples. Empirically, we demonstrate significant improvements over standard retrieval methods across a suite of classification tasks, highlighting the benefits of structure-aware, diverse example selection for ICL in real-world, label-scarce scenarios.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Commonsense Reasoning | HellaSwag | HellaSwag Accuracy71.02 | 897 | |
| Sentiment Classification | SST-2 | Accuracy94.28 | 220 | |
| Natural Language Inference | QNLI | Accuracy71.68 | 93 | |
| Paraphrase Detection | MRPC | Accuracy75.27 | 90 | |
| Sentiment Classification | SST-5 | Accuracy49.59 | 73 | |
| Classification | SST-5, MRPC, QNLI, SWAG, IMDB, DBPedia, BoolQ OpenAI large text embeddings (test) | Accuracy83.5 | 16 | |
| Question Answering and Reasoning | NQ, TriviaQA, HotpotQA, and GSM8K Average | Average Accuracy72.5 | 16 | |
| Commonsense Question Answering | CMSQA | Accuracy72.89 | 15 |