Low-Confidence Gold: Refining Low-Confidence Samples for Efficient Instruction Tuning
The effectiveness of instruction fine-tuning for Large Language Models is fundamentally constrained by the quality and efficiency of training datasets. This work introduces Low-Confidence Gold (LCG), a novel filtering framework that employs centroid-based clustering and confidence-guided selection for identifying valuable instruction pairs. Through a semi-supervised approach using a lightweight classifier trained on representative samples, LCG curates high-quality subsets while preserving data diversity. Experimental evaluation demonstrates that models fine-tuned on LCG-filtered subsets of 6K samples achieve superior performance compared to existing methods, with substantial improvements on MT-bench and consistent gains across comprehensive evaluation metrics. The framework's efficacy while maintaining model performance establishes a promising direction for efficient instruction tuning.
Code (0)
등록된 구현이 없습니다.
Tasks
DiversitySimilar Papers 제목 키워드 기반
Learning Robust Pair Confidence for Multimodal Emotion-Cause Pair Extraction
Multimodal emotion-cause pair extraction (MECPE) requires reliable pair confidence over candidate pairs. Existing pair scorers commonly use pair-level cross entropy over valid candidates, which treats links mostly indepe…
Emotion-Cause Pair ExtractionTypicalness-Aware Learning for Failure Detection
Deep neural networks (DNNs) often suffer from the overconfidence issue, where incorrect predictions are made with high confidence scores, hindering the applications in critical systems. In this paper, we propose a novel …
Deconstructing the Goldilocks Zone of Neural Network Initialization
The second-order properties of the training loss have a massive impact on the optimization dynamics of deep learning models. Fort & Scherlis (2019) discovered that a large excess of positive curvature and local convexity…
Unleashing the Potential of All Test Samples: Mean-Shift Guided Test-Time Adaptation
Visual-language models (VLMs) like CLIP exhibit strong generalization but struggle with distribution shifts at test time. Existing training-free test-time adaptation (TTA) methods operate strictly within CLIP's original …
Test-time AdaptationAre You Sure You're Sure? On the Impact of Instruction Tuning on Confidence and Lexical Diversity
Instruction-tuned language models achieve strong performance across a range of generation tasks, but have also recently been shown to exhibit verbalized overconfidence. In question answering, verbalized model overconfide…
Question AnsweringAnswer Selection