arXiv subjects

Mario Sanz-Guerrero

Publications and source records attributed to Mario Sanz-Guerrero.

11 recordsLinked to original sources

arXiv2026

Dating the Model: Hidden Dates in System Prompts Affect LLM Evaluation

Reproducibility is essential for scientific research, yet prior work shows that LLM outputs vary with hardware and batching. We identify an overlooked factor: the hidden injection of the current date into system prompts, which users cannot control and which changes every day. Across 9 recent LLMs and 6 datasets spanning multiple-choice QA (MCQA), math reasoning, code generation, and machine translation, performance varies solely with the current date, with deltas of up to 6% on MCQA, 14% on math reasoning, 7% on code generation, and 2.84 BLEU on machine translation. Model rankings also shift, affecting leaderboards. This date effect exceeds other sources of non-determinism, such as batch size and numerical precision. Standard prompting techniques -- chain-of-thought and few-shot prompting -- do not reduce the sensitivity; chain-of-thought even amplifies it. Our findings underscore the need for careful evaluation protocols to ensure reproducibility and fair comparisons in LLM research.

cs.CL↗

arXiv2026

Calibration as a First-Class Criterion in LLM Evaluation

Calibration of language models -- the alignment between expressed or implicit confidence and empirical correctness -- is a well-studied subfield within NLP. Methods to measure it already exist. The problem is adoption: outside this subfield, NLP research regularly introduces new models, datasets, and benchmarks without checking whether the model's confidence scores are meaningful. We argue that this adoption gap is a major obstacle to trustworthy LLM evaluation. Miscalibration causes problems in two distinct areas: at deployment, where overconfident mistakes cause real harm, and inside the research pipeline, where methods like LLM-as-a-judge, synthetic data generation, and active learning rely on calibrated confidence without verifying it. Standard calibration metrics only require two inputs per example: a confidence score and a correctness judgment. Most benchmarks in use today already provide both, meaning calibration can be reported immediately. For open-ended generation, however, defining these two inputs is still an open challenge. We argue that each NLP subfield should pair its main performance metric with a calibration score and call for treating calibration as an essential property of every model rather than a niche topic.

cs.CL↗

arXiv2026

Perspectives on Cross-Lingual Consistency in LLMs for Medical Questions

Should multilingual LLMs answer medical questions consistently across input languages, or adapt responses to cultural cues? Existing multilingual medical benchmarks usually assume that medically correct answers should remain consistent across languages and treat cross-lingual variation as model error. In contrast, cultural adaptation research argues that appropriate medical answers may legitimately differ across contexts. We review the multilingual medical NLP literature through these two perspectives, we identify three gaps: limited stakeholder perspectives (e.g., of medical professionals), a lack of empirical evidence on which approach better serves users, and no benchmarks capable of distinguishing universally correct from culture-specific cases. To address the first gap, we survey 356 participants across three stakeholder groups (medical, NLP, and anthropology professionals) in three countries (Germany, Spain, and the United States). Anthropologists consistently favor adaptation, while medical and NLP respondents remain divided, with notable divergence between U.S. and European medical professionals. LLMs prompted with profession and country personas fail to reproduce this variation, overestimating cross-lingual consistency preference among NLP and medical personas. We conclude that neither consistency nor adaptation can currently be considered clearly preferable, highlighting the need for empirical evidence on which approach better serves users across cultural contexts.

cs.CL↗

arXiv2026

Large Language Models Are Overconfident in Their Own Responses

Prior work has shown that instruction-tuned large language models (LLMs) are less well calibrated than their base pre-trained counterparts. However, little is known about the frequently used chat template's effect on the calibration of conversational LLMs. In this work, we investigate the mechanisms driving this miscalibration by decoupling the effects of the post-training algorithm and the chat format. We find that, while instruction tuning fundamentally harms calibration, the chat template aggravates the issue through an "ownership bias" -- models are significantly more confident in their own answers than in identical answers provided by a user. Extensive experiments across six recent open-weight LLMs, three benchmarks, and three confidence elicitation methods show that models assign up to 26% higher confidence to their own responses. Leveraging this insight, we propose a simple inference-time strategy: framing the model's answer as user input during confidence elicitation. This approach significantly reduces overconfidence and improves calibration by up to 26% without the need for retraining, narrowing the gap between base and instruction-tuned models.

cs.CL↗

arXiv2025

NALA_MAINZ at BLP-2025 Task 2: A Multi-agent Approach for Bangla Instruction to Python Code Generation

This paper presents JGU Mainz's winning system for the BLP-2025 Shared Task on Code Generation from Bangla Instructions. We propose a multi-agent-based pipeline. First, a code-generation agent produces an initial solution from the input instruction. The candidate program is then executed against the provided unit tests (pytest-style, assert-based). Only the failing cases are forwarded to a debugger agent, which reruns the tests, extracts error traces, and, conditioning on the error messages, the current program, and the relevant test cases, generates a revised solution. Using this approach, our submission achieved first place in the shared task with a $Pass@1$ score of 95.4. We also make our code public.

cs.CL↗

arXiv2025

Mitigating Label Length Bias in Large Language Models

Large language models (LLMs) are powerful zero- and few-shot learners. However, when predicting over a set of candidate options, LLMs suffer from label biases, and existing calibration methods overlook biases arising from multi-token class labels. We tackle an issue we call label length bias, where labels of different lengths are treated inconsistently, even after standard length normalization. To mitigate it, we propose normalized contextual calibration (NCC), an effective method that normalizes and calibrates predictions at the full-label level. NCC achieves statistically significant improvements over prior approaches across multiple datasets and models, with gains of up to 10% F1. Moreover, NCC extends bias mitigation to broader tasks such as multiple-choice question answering. Our analysis shows that, when combined with in-context learning, NCC is less sensitive to few-shot example selection, requires fewer examples for competitive performance, and produces more reliable confidence estimates. These findings highlight the importance of mitigating full-label biases to improve the performance and robustness of LLM-based methods, particularly in real-world applications where class labels naturally consist of multiple tokens.

cs.CL↗

arXiv2025

JGU Mainz's Submission to the WMT25 Shared Task on LLMs with Limited Resources for Slavic Languages: MT and QA

This paper presents the JGU Mainz submission to the WMT25 Shared Task on LLMs with Limited Resources for Slavic Languages: Machine Translation and Question Answering, focusing on Ukrainian, Upper Sorbian, and Lower Sorbian. For each language, we jointly fine-tune a Qwen2.5-3B-Instruct model for both tasks with parameter-efficient finetuning. Our pipeline integrates additional translation and multiple-choice question answering (QA) data. For Ukrainian QA, we further use retrieval-augmented generation. We also apply ensembling for QA in Upper and Lower Sorbian. Experiments show that our models outperform the baseline on both tasks.

cs.CL↗

arXiv2025

Mind the Gap: A Closer Look at Tokenization for Multiple-Choice Question Answering with LLMs

When evaluating large language models (LLMs) with multiple-choice question answering (MCQA), it is common to end the prompt with the string "Answer:" to facilitate automated answer extraction via next-token probabilities. However, there is no consensus on how to tokenize the space following the colon, often overlooked as a trivial choice. In this paper, we uncover accuracy differences of up to 11% due to this (seemingly irrelevant) tokenization variation as well as reshuffled model rankings, raising concerns about the reliability of LLM comparisons in prior work. Surprisingly, we are able to recommend one specific strategy -- tokenizing the space together with the answer letter -- as we observe consistent and statistically significant performance improvements. Additionally, it improves model calibration, enhancing the reliability of the model's confidence estimates. Our findings underscore the importance of careful evaluation design and highlight the need for standardized, transparent evaluation protocols to ensure reliable and comparable results.

cs.CL↗

arXiv2025

Credit Risk Meets Large Language Models: Building a Risk Indicator from Loan Descriptions in P2P Lending

Peer-to-peer (P2P) lending connects borrowers and lenders through online platforms but suffers from significant information asymmetry, as lenders often lack sufficient data to assess borrowers' creditworthiness. This paper addresses this challenge by leveraging BERT, a Large Language Model (LLM) known for its ability to capture contextual nuances in text, to generate a risk score based on borrowers' loan descriptions using a dataset from the Lending Club platform. We fine-tune BERT to distinguish between defaulted and non-defaulted loans using the loan descriptions provided by the borrowers. The resulting BERT-generated risk score is then integrated as an additional feature into an XGBoost classifier used at the loan granting stage, where decision-makers have limited information available to guide their decisions. This integration enhances predictive performance, with improvements in balanced accuracy and AUC, highlighting the value of textual features in complementing traditional inputs. Moreover, we find that the incorporation of the BERT score alters how classification models utilize traditional input variables, with these changes varying by loan purpose. These findings suggest that BERT discerns meaningful patterns in loan descriptions, encompassing borrower-specific features, specific purposes, and linguistic characteristics. However, the inherent opacity of LLMs and their potential biases underscore the need for transparent frameworks to ensure regulatory compliance and foster trust. Overall, this study demonstrates how LLM-derived insights interact with traditional features in credit risk modeling, opening new avenues to enhance the explainability and fairness of these models.

q-fin.RM↗

arXiv2025

Corrective In-Context Learning: Evaluating Self-Correction in Large Language Models

In-context learning (ICL) has transformed the use of large language models (LLMs) for NLP tasks, enabling few-shot learning by conditioning on labeled examples without finetuning. Despite its effectiveness, ICL is prone to errors, especially for challenging examples. With the goal of improving the performance of ICL, we propose corrective in-context learning (CICL), an approach that incorporates a model's incorrect predictions alongside ground truth corrections into the prompt, aiming to enhance classification accuracy through self-correction. However, contrary to our hypothesis, extensive experiments on text classification tasks demonstrate that CICL consistently underperforms standard ICL, with performance degrading as the proportion of corrections in the prompt increases. Our findings indicate that CICL introduces confusion by disrupting the model's task understanding, rather than refining its predictions. Additionally, we observe that presenting harder examples in standard ICL does not improve performance, suggesting that example difficulty alone may not be a reliable criterion for effective selection. By presenting these negative results, we provide important insights into the limitations of self-corrective mechanisms in LLMs and offer directions for future research.

cs.CL↗

arXiv2025

Asking Again and Again: Exploring LLM Robustness to Repeated Questions

This study investigates whether repeating questions within prompts influences the performance of large language models (LLMs). We hypothesize that reiterating a question within a single prompt might enhance the model's focus on key elements of the query. We evaluate five recent LLMs -- including GPT-4o-mini, DeepSeek-V3, and smaller open-source models -- on three reading comprehension datasets under different prompt settings, varying question repetition levels (1, 3, or 5 times per prompt). Our results demonstrate that question repetition can increase models' accuracy by up to $6\%$. However, across all models, settings, and datasets, we do not find the result statistically significant. These findings provide insights into prompt design and LLM behavior, suggesting that repetition alone does not significantly impact output quality.

cs.CL↗