On the Relation between Sensitivity and Accuracy in In-context Learning
In-context learning (ICL) suffers from oversensitivity to the prompt, making it unreliable in real-world scenarios. We study the sensitivity of ICL with respect to multiple perturbation types. First, we find that label bias obscures the true sensitivity, and therefore prior work may have significantly underestimated ICL sensitivity. Second, we observe a strong negative correlation between ICL sensitivity and accuracy: predictions sensitive to perturbations are less likely to be correct. Motivated by these findings, we propose \textsc{SenSel}, a few-shot selective prediction method that abstains from sensitive predictions. Experiments on ten classification datasets show that \textsc{SenSel} consistently outperforms two commonly used confidence-based and entropy-based baselines on abstention decisions.
Code (1)
Tasks
In-Context LearningRelationSensitivitySimilar Papers 제목 키워드 기반
Controllable Context Sensitivity and the Knob Behind It
When making predictions, a language model must trade off how much it relies on its context vs. its prior knowledge. Choosing how sensitive the model is to its context is a fundamental functionality, as it enables the mod…
Question AnsweringRetrieval-augmented GenerationSensitivityBetween Rules and Reality: On the Context Sensitivity of LLM Moral Judgment
A human's moral decision depends heavily on the context. Yet research on LLM morality has largely studied fixed scenarios. We address this gap by introducing Contextual MoralChoice, a dataset of moral dilemmas with syste…
Gradient-based Active Learning with Gaussian Processes for Global Sensitivity Analysis
Global sensitivity analysis of complex numerical simulators is often limited by the small number of model evaluations that can be afforded. In such settings, surrogate models built from a limited set of simulations can s…
Gaussian ProcessesActive LearningHow are Prompts Different in Terms of Sensitivity?
In-context learning (ICL) has become one of the most popular learning paradigms. While there is a growing body of literature focusing on prompt engineering, there is a lack of systematic analysis comparing the effects of…
In-Context LearningPrompt EngineeringSensitivityYour spouse needs professional help: Determining the Contextual Appropriateness of Messages through Modeling Social Relationships
Understanding interpersonal communication requires, in part, understanding the social context and norms in which a message is said. However, current methods for identifying offensive content in such communication largely…