# Task Complexity Matters: An Empirical Study of Reasoning in LLMs for Sentiment Analysis

arXiv: https://arxiv.org/abs/2602.24060 · HTML: https://mlsanity.com/paper/2602.24060/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Audio-language model reasoning, one of 17 directions in LLM reasoning.
- Builds on 8 of its 18 references that are in the map.
- First posted on arXiv 27 Feb 2026.

## Placed in

- LLM reasoning / [Audio-language model reasoning](https://mlsanity.com/lr/audio-language-model-reasoning/)

## Builds on (most cited first, 8 of 8)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes](https://mlsanity.com/paper/2305.02301/) (2023)
- [A Survey on Knowledge Distillation of Large Language Models](https://mlsanity.com/paper/2402.13116/) (2024)
- [Sentiment Analysis in the Era of Large Language Models: A Reality Check](https://mlsanity.com/paper/2305.15005/) (2023)
- [Deciphering the Factors Influencing the Efficacy of Chain-of-Thought: Probability, Memorization, and Noisy Reasoning](https://mlsanity.com/paper/2407.01687/) (2024)
- [Reasoning or Overthinking: Evaluating Large Language Models on Financial Sentiment Analysis](https://mlsanity.com/paper/2506.04574/) (2025)
- [Explainable Sentiment Analysis with DeepSeek-R1: Performance, Efficiency, and Few-Shot Learning](https://mlsanity.com/paper/2503.11655/) (2025)
- [Targeted Distillation for Sentiment Analysis](https://mlsanity.com/paper/2503.03225/) (2025)

## Built on by (most cited first, 2 of 2)

- [Curating Merchant-Matching Training Data with Two Confidence-Gated Local LLM Judges](https://mlsanity.com/paper/2609.33878/) (2026)
- [QC-Stark: A Multi-Task Benchmark Revealing Capability Dissociations in LLMs Evaluated on Quantum Computing Tasks](https://mlsanity.com/paper/2609.35581/) (2026)

## Abstract (arXiv)

Large language models (LLMs) with reasoning capabilities have fueled a compelling narrative that reasoning universally improves performance across language tasks. We test this claim through a comprehensive evaluation of 504 configurations across seven model families--including adaptive, conditional, and reinforcement learning-based reasoning architectures--on sentiment analysis datasets of varying granularity (binary, five-class, and 27-class emotion). Our findings reveal that reasoning effectiveness is strongly task-dependent, challenging prevailing assumptions: (1) Reasoning shows task-complexity dependence--binary classification degrades up to -19.9 F1 percentage points (pp), while 27-class emotion recognition gains up to +16.0pp; (2) Distilled reasoning variants underperform base models by 3-18 pp on simpler tasks, though few-shot prompting enables partial recovery; (3) Few-shot learning improves over zero-shot in most cases regardless of model type, with gains varying by architecture and task complexity; (4) Pareto frontier analysis shows base models dominate efficiency-performance trade-offs, with reasoning justified only for complex emotion recognition despite 2.1x-54x computational overhead. We complement these quantitative findings with qualitative error analysis revealing that reasoning degrades simpler tasks through systematic over-deliberation, offering mechanistic insight beyond the high-level overthinking hypothesis.