Related papers: FormInv: A Measurement Protocol for Semantic Invar…

Large Language Models (LLMs) often exhibit inconsistent behavior when answering paraphrased questions, suggesting a reliance on surface-level patterns rather than true semantic understanding. To address this limitation, we introduce RoParQ,…

As large language models (LLMs) are increasingly deployed in high-stakes applications, robust uncertainty estimation is essential for ensuring the safe and trustworthy deployment of LLMs. We present the most comprehensive study to date of…

Large language models have been shown to behave inconsistently in response to meaning-preserving paraphrastic inputs. At the same time, researchers evaluate the knowledge and reasoning abilities of these models with test evaluations that do…

Safety-aligned language models often refuse prompts that are actually harmless. Current evaluations mostly report global rates such as false rejection or compliance. These scores treat each prompt alone and miss local inconsistency, where a…

We conduct a systematic audit of three widely used reasoning benchmarks, SocialIQa, FauxPas-EAI, and ToMi, and uncover pervasive flaws in both benchmark items and evaluation methodology. Using five LLMs (GPT-{3, 3.5, 4, o1}, and LLaMA 3.1)…

Large language models (LLMs) have demonstrated impressive performance on reasoning tasks, including mathematical reasoning. However, the current evaluation mostly focuses on carefully constructed benchmarks and neglects the consideration of…

Evaluating statement autoformalization, translating natural language mathematics into formal languages like Lean 4, remains a significant challenge, with few metrics, datasets, and standards to robustly measure progress. In this work, we…

Large language models must satisfy hard orthographic constraints during controlled text generation, yet systematic cross-family evaluation remains limited. We evaluate 39 configurations spanning three model families (Qwen3, Claude Haiku…

LLMs have achieved remarkable success in complex reasoning tasks, yet current evaluation approaches predominantly rely on final-answer correctness, offering limited insight into the underlying reasoning processes that produce those answers.…

Radiology report generation (RRG) has shown great potential in assisting radiologists by automating the labor-intensive task of report writing. While recent advancements have improved the quality and coherence of generated reports, ensuring…

Guard models are a critical component of LLM safety, but their sensitivity to superficial linguistic variations remains a key vulnerability. We show that even meaning-preserving paraphrases can cause large fluctuations in safety scores,…

We introduce CRYSTAL (Clear Reasoning via Yielded Steps, Traceability, and Logic), a diagnostic benchmark with 6,372 instances that evaluates multimodal reasoning through verifiable intermediate steps. We propose two complementary metrics:…

Composed Image Retrieval (CIR) has made significant progress, yet current benchmarks are limited to single ground-truth answers and lack the annotations needed to evaluate false positive avoidance, robustness and multi-image reasoning. We…

Automatic evaluation remains an open research question in Natural Language Generation. In the context of Sentence Simplification, this is particularly challenging: the task requires by nature to replace complex words with simpler ones that…

Formal mathematical reasoning remains a critical challenge for artificial intelligence, hindered by limitations of existing benchmarks in scope and scale. To address this, we present FormalMATH, a large-scale Lean4 benchmark comprising…

Mathematical reasoning models are widely deployed in education, automated tutoring, and decision support systems despite exhibiting fundamental computational instabilities. We demonstrate that state-of-the-art models (Qwen2.5-Math-7B)…

LLMs deployed for natural-language querying of analytical databases suffer from two intertwined failures - incorrect answers and confident hallucinations - both rooted in the same cause: the model is forced to infer business semantics that…

In this paper, we present our system for SemEval-2026 Task 6 (CLARITY) on response clarity and evasion detection in question-answer pairs from U.S. presidential interviews, comparing fine-tuned encoders with prompt-based LLMs. Our LLM…

Position-controlled evaluation is standard for retrieval tasks such as Needle-in-a-Haystack and RULER, but mainstream reasoning benchmarks do not control positional placement of target tasks in long contexts. We audit 11 long-context…

Qualitative analysis plays a pivotal role in understanding the human and social aspects of software engineering. However, it remains a demanding process shaped by the subjective interpretation of individual researchers and sensitive to…