Logit Separability-Driven Samples and Multiple Class-Related Words Selection for Advancing In-Context Learning
This work addresses the challenge of improving response quality in in-context learning for large language models, though it is incremental as it builds on existing demonstration optimization methods.
The paper tackled the problem of organizing in-context learning demonstrations for large language models by introducing logit separability to assess sample and label clarity, and incorporating multiple class-related words per sample, resulting in significant performance improvements across seven classification datasets.
Effective organization of in-context learning (ICL) demonstrations is key to improving the quality of large language model (LLM) responses. To create better sample-label pairs that instruct LLM understanding, we introduce logit separability, a criterion to assess the clarity of both samples and class-related words at the logit level. This facilitates the optimization of sample and label selection, enhancing the precision of information provided in ICL demonstrations. Additionally, we find that incorporating multiple class-related words for each sample, rather than relying on a single class name, improves performance by offering a broader range of label information. Building on these insights, we propose LICL, a logit separability-based method that jointly organizes samples and integrates multiple class-related words into each sample-label pair. Evaluations across seven classification datasets show that this approach significantly improves ICL performance by providing clearer instructions and richer label information.