Jul 2026· Applied and Computational Engineering· 0 citations
TL;DR
Overall, LLMs offer a flexible and annotation-light alternative to traditional fine-tuning, but their successful deployment in finance will depend on robust prompt engineering and solid factual grounding.
Abstract
Financial sentiment analysis has long relied on labeled data to fine-tune models like FinBERT, a process that is both costly and time-consuming. The arrival of large language models (LLMs) has changed the landscape: with zero-shot and few-shot prompting, one can now extract sentiment from financial texts using few or no annotated examples. This survey takes stock of how LLMs are being applied to this task. It begins by clarifying the core ideas behind in-context learning and chain-of-thought prompting. It then examines a range of prompt designs that have been developed to cope with the peculiarities of financial writing, such as numerical expressions, implicit sentiment and long documents. A comparison of LLM performance on standard benchmarks against fine-tuned domain models shows that general-purpose LLMs are often competitive, especially when prompts are carefully crafted. Yet three problems remain unresolved: numerical reasoning errors, hallucination, and the practical hurdles of cost, latency, and privacy. These challenges are discussed in detail, and retrieval-augmented generation, trustworthiness frameworks, and efficient open-source models are pointed out as the most promising paths forward. Overall, LLMs offer a flexible and annotation-light alternative to traditional fine-tuning, but their successful deployment in finance will depend on robust prompt engineering and solid factual grounding.
— By comparing and analysing the performance of several machine learning algorithms on fine-grained sentiment classification problems to examine their suitability and shortcomings for use as models in sentiment analysis. Sentiment analysis remains a prominent research area in Natural Language Processing (NLP). However, systematic comparisons of whether these methods demonstrate superiority in fine-grained classification settings have yet to be conducted. Based on the GoEmotions dataset that includes 28 sentiment classes as our experiment’s reference for evaluating four kinds of models: logistic regression, BiLSTM, BERT, and the large-scale language model Claude (claude-haiku-4-5). The experiments employ TF-IDF feature extraction, sequence encoding, pre-training with fine-tuning, and zero-shot and few-shot prompting strategies. Based on this experiment, the model’s generalisation capabilities improve as architectural depth increases. The accuracy of logistic regression was only 55.11%; BERT obtained the highest F1 score and performed the best overall, and bidirectional pre-trained representations were most valuable among other methods. However, large language models perform significantly worse on the 28-class classification task in zero-shot settings, suggesting that they are better suited for generative and open-ended emotional interaction than for standardized classification benchmarks.
Shangjiafeng Guo· International journal of eng...· 0 citations
The findings show that QLoRA is effective for financial sentiment adaptation, while also documenting a clear gap between classification accuracy and tradable cross-sectional signals.
The work provides a reproducible, explainable, operationally applicable model of sentiment analysis in operationally sensitive, high-stakes Twitter sentiment analysis, and validate the hypothesis that hybrid stacking is an effective method for leveraging the complementary nature of lexical and contextual representations.
D. Abate, Nilay Mistry· International Research Journ...· 0 citations
Large language models (LLMs) prompt a recurring question in research computing: should classical methods like Naive Bayes (NB) be retired? We benchmark Complement Naive Bayes against zero-shot and few-shot LLMs spanning four model families and a 37 × range in scale (27B to a 1T-parameter mixture-of-experts) across text classification tasks. LLMs dominate only in zero-data regimes (98.0% vs 88.2% on Amazon Polarity sentiment), and even that win is contamination-prone: on a low-contamination sentiment task NB beats the zero-shot LLM (81.7% vs 73.0%). However, once labeled data is available (e.g., AG News), NB reaches 89.1% accuracy, statistically indistinguishable from the zero-shot 27B LLM (89.0%) and better than the 397B frontier model (84.8%), at thousands of samples/sec on a commodity CPU. Fine-tuned DistilBERT reaches 90.6% but at far lower throughput than NB at batch size 1 (Table 2). Our measured GPU throughput analysis shows small-LLM batched inference is 40–486 × slower than NB CPU inference (the multiplier depends strongly on the host CPU), exposing a structural gap bounded by memory bandwidth, with roughly two orders of magnitude lower energy per sample. For resource-constrained HPC practitioners performing text classification with labeled data, NB remains the optimal choice. We show the decision line is task-dependent (NB reaches LLM parity around N ∼ 104 labels for topic classification, while zero-data sentiment favors the LLM at all N tested) and provide a Kubernetes Helm operator that automates model selection using configurable thresholds and verifiable Prometheus metrics.
M. Sada, John J. Graham, M. Tatineni et al.· Practice and Experience in A...· 0 citations
Social media sentiment analysis has become one of the most significant instruments for understanding the opinion of the population in the spheres of healthcare, politics, and education. Yet, large language models (LLMs) remain unevenly distributed in their linguistic coverage, failing to adequately serve a large portion of the world's languages. This study evaluates five state-of-the-art LLMs: GPT-4o, Gemini 2.0 Flash, DeepSeek-V3, Mistral Large, and Claude 3.7 Sonnet on three-class sentiment classification across 36 datasets spanning 36 languages, with emphasis on low- and medium-resource settings, using zero-shot and few-shot prompting without task-specific fine-tuning. In addition to the traditional measures of performance per language, the study presents a hierarchical analysis of languages based on a genealogical tree of Indo-European, Afro-Asiatic, Niger-Congo, Turkic, Austronesian, and English Creole language families, so that it is possible to identify the systematic patterns of performance superiority and inferiority among the language families. The findings show that few-shot prompting improves the results of a vast majority of languages, with several models approaching or surpassing the performance of the state-of-the-art benchmark of task-specific models. The GPT-4o and Claude achieved the highest performance in the high-resource and medium-resource settings, and Gemini is a competent trade-off that allows balancing the performance and the computational cost. Although it has lower zero-shot performance, Mistral benefits the most from few-shot prompting and becomes highly competitive in the few-shot setting. Despite these developments, the level of performance on low-resource languages, such as Oromo, Xitsonga, Azerbaijani, and Twi, remains significantly lower, underscoring that progress in multilingual LLMs requires moving beyond English-centric evaluation toward genuinely representative and globally inclusive benchmarks.
Muhamet Kastrati, Abdul Manaf, A. Imran et al.· Frontiers in Artificial Inte...· 0 citations
