Emo-Jev: Probabilistic Reasoning for Emotion Classification with Jev

查看原文 / View Original 编辑 / Edit

Computer Science > Computation and Language arXiv:2610.08829 (cs) [Submitted on 27 Sep 2026] Title: Emo-Jev: Probabilistic Reasoning for Emotion Classification with Jev Authors: Yazhou Zhang, Junhao Yu.

计算机科学 > 计算与语言 arXiv:2610.08829 (cs) [提交于 2026 年 9 月 27 日] 标题：Emo-Jev：基于 Jev 的情感分类概率推理作者：Yazhou Zhang, Junhao Yu。

Abstract: Jev offers an alternative interface for language understanding: given an input and predefined questions, it returns probabilistic decisions rather than free-form responses. Whether this interface can support effective reasoning for text classification against leading LLMs remains an open question.

摘要：Jev 为语言理解提供了一种替代接口：给定输入和预定义问题，它返回的是概率决策而非自由形式的回答。该接口是否能支持针对领先大语言模型（LLM）的有效文本分类推理，仍是一个悬而未决的问题。

We introduce Emo-Jev, a training-free framework with two complementary implementations. Emo-Jev-D decomposes classification into task-specific atomic judgments and composes their probabilities into a final prediction. Emo-Jev-SC constructs multiple judgment paths from complementary perspectives and aggregates their predictions into a consensus decision.

我们引入了 Emo-Jev，这是一个无需训练的框架，包含两种互补的实现方式。Emo-Jev-D 将分类任务分解为特定任务的原子判断，并将它们的概率组合成最终预测。Emo-Jev-SC 从互补的角度构建多个判断路径，并将它们的预测汇总为共识决策。

We evaluate Emo-Jev on eight datasets spanning sentiment analysis, emotion recognition, sarcasm detection and humor detection, comparing against direct Jev classification and five SoTA LLMs under input/output and chain-of-thought reasoning. Standard Jev achieves 62.93% average macro-F1 versus 67.28% for the strongest LLM baseline, with lower observed latency and generally lower cost.

我们在涵盖情感分析、情绪识别、讽刺检测和幽默检测的八个数据集上评估了 Emo-Jev，并将其与直接的 Jev 分类以及在输入/输出和思维链推理下的五个最先进（SoTA）大语言模型进行了比较。标准 Jev 达到了 62.93% 的平均宏观 F1 值，而最强的大语言模型基准为 67.28%，且前者观察到的延迟更低，成本通常也更低。