LLMs Show Surface-Form Brittleness Under Paraphrase Stress Tests

• URL: http://arxiv.org/abs/2510.08616v1
• Date: Wed, 08 Oct 2025 06:04:33 GMT
• Title: LLMs Show Surface-Form Brittleness Under Paraphrase Stress Tests
• Abstract summary: Benchmark scores for Large Language Models (LLMs) can be inflated by memorization of test items or near duplicates.<n>We present a protocol that probes generalization by re-evaluating models on paraphrased versions of benchmark questions.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Benchmark scores for Large Language Models (LLMs) can be inflated by memorization of test items or near duplicates. We present a simple, protocol that probes generalization by re-evaluating models on paraphrased versions of benchmark questions. Using Mistral-7B-Instruct and Qwen2.5-7B-Instruct, we measure the accuracy gap between original and paraphrased items on ARC-Easy and ARC-Challenge. Our pipeline controls decoding, enforces multiple-choice output format, and includes a robust paraphrase-cleaning step to preserve semantics. We find that paraphrasing induces a non-trivial accuracy drop (original vs. paraphrased), consistent with prior concerns about contamination and brittle surface-form shortcuts.

Related papers

• Structuring Semantic Embeddings for Principle Evaluation: A Prototype-Guided Contrastive Learning Approach [58.15427952222424]
  General-purpose text embeddings are widely deployed for such tasks, but broad semantic similarity can place semantically similar yet task-distinct examples.<n>We introduce Prototype-Guided Contrastive Learning (PGCL), a prototype-guided geometric regularization module built on top of frozen text embeddings.
  arXiv  Detail & Related papers  (2026-08-15T13:19:25Z)
• SENSE: Semantic Embedding Navigation with Soft-gated Evaluation for Retrieval-based Speculative Decoding [6.448291801598507]
  Speculative Decoding (SD) accelerates Large Language Model (LLM) inference.<n>RSD is favored for its plug-and-play versatility, but its potential is impeded by rigid lexical dependencies.<n>SENSE establishes robust semantic alignment, which empowers the Soft-gated Evaluation module to validate semantic equivalence.
  arXiv  Detail & Related papers  (2026-04-14T04:17:03Z)
• LIBERO-Para: A Diagnostic Benchmark and Metrics for Paraphrase Robustness in VLA Models [15.955490895662384]
  We introduce LIBERO-Para, a benchmark for fine-grained analysis of linguistic generalization.<n>We observe consistent performance degradation of 22-52 pp under paraphrasing.<n>We propose PRIDE, a metric that quantifies paraphrase difficulty using semantic and syntactic factors.
  arXiv  Detail & Related papers  (2026-03-30T11:27:34Z)
• Beyond Raw Detection Scores: Markov-Informed Calibration for Boosting Machine-Generated Text Detection [105.14032334647932]
  Machine-generated texts (MGTs) pose risks such as disinformation and phishing, highlighting the need for reliable detection.<n> Metric-based methods, which extract statistically distinguishable features of MGTs, are often more practical than complex model-based methods that are prone to overfitting.<n>We propose a Markov-informed score calibration strategy that models two relationships of context detection scores that may aid calibration.
  arXiv  Detail & Related papers  (2026-02-08T16:06:12Z)
• Residual Context Diffusion Language Models [90.07635240595926]
  Residual Context Diffusion (RCD) is a module that converts discarded token representations into contextual residuals and injects them back for the next denoising step.<n>RCD consistently improves frontier dLLMs by 5-10 points in accuracy with minimal extra computation overhead.
  arXiv  Detail & Related papers  (2026-01-30T13:16:32Z)
• DVD: A Robust Method for Detecting Variant Contamination in Large Language Model Evaluation [24.086354908256293]
  textbfDVD is a single-sample detector that models the local output distribution induced by temperature sampling.<n>We construct the first benchmark for variant contamination across two domains Omni-MATH and SuperGPQA.<n>textbfDVD consistently outperforms perplexity-based, Min-$k$%++, edit-distance (CDD), and embedding-similarity baselines.
  arXiv  Detail & Related papers  (2026-01-08T12:48:40Z)
• RoParQ: Paraphrase-Aware Alignment of Large Language Models Towards Robustness to Paraphrased Questions [0.0]
  Large Language Models (LLMs) often exhibit inconsistent behavior when answering paraphrased questions.<n>We introduce RoParQ, a benchmark to evaluate cross-paraphrase consistency in closed-book multiple-choice QA.<n>We also propose XParaCon, a novel evaluation metric that quantifies a model's robustness.
  arXiv  Detail & Related papers  (2025-11-26T16:40:53Z)
• Distinguishing Repetition Disfluency from Morphological Reduplication in Bangla ASR Transcripts: A Novel Corpus and Benchmarking Analysis [0.0]
  We introduce the first publicly available, 20,000-row Bangla corpus, manually annotated to explicitly distinguish between these two phenomena in noisy ASR transcripts.<n>We benchmark this novel resource using two paradigms: state-of-the-art multilingual Large Language Models (LLMs) and task-specific fine-tuning of encoder models.
  arXiv  Detail & Related papers  (2025-11-17T09:06:01Z)
• Do LLMs Know They Are Being Tested? Evaluation Awareness and Incentive-Sensitive Failures in GPT-OSS-20B [1.948261185683419]
  We investigate whether "evaluation scent" inflates measured performance without commensurate capability gains.<n>We run six paired A/B scenarios that hold task content and decoding fixed while varying framing.<n>We provide a reproducible A/B framework (prompt banks, validators, per-run scores, scripts) and practical guidance.
  arXiv  Detail & Related papers  (2025-10-08T09:49:05Z)
• DRES: Benchmarking LLMs for Disfluency Removal [27.083825614818135]
  Disfluencies, such as "um," "uh," interjections, parentheticals, and edited statements, remain a persistent challenge for speech-driven systems.<n>Disfluency Removal Evaluation Suite, a controlled text-level benchmark, establishes a reproducible semantic upper bound for this task.<n>DRES builds on human-annotated Switchboard transcripts, isolating disfluency removal from ASR errors and acoustic variability.
  arXiv  Detail & Related papers  (2025-09-24T17:08:12Z)
• CompassVerifier: A Unified and Robust Verifier for LLMs Evaluation and Outcome Reward [50.97588334916863]
  We develop CompassVerifier, an accurate and robust lightweight verifier model for evaluation and outcome reward.<n>It demonstrates multi-domain competency spanning math, knowledge, and diverse reasoning tasks, with the capability to process various answer types.<n>We introduce VerifierBench benchmark comprising model outputs collected from multiple data sources, augmented through manual analysis of metaerror patterns to enhance CompassVerifier.
  arXiv  Detail & Related papers  (2025-08-05T17:55:24Z)
• Explanation based In-Context Demonstrations Retrieval for Multilingual Grammatical Error Correction [19.95974494301433]
  Grammatical error correction (GEC) aims to correct grammatical, spelling, and semantic errors in natural language text.<n>We propose a novel retrieval method based on natural language grammatical error explanations (GEE)<n>Our method retrieves suitable few-shot demonstrations by matching the GEE of the test input with that of pre-constructed database samples.
  arXiv  Detail & Related papers  (2025-02-12T15:41:43Z)
• Localizing Factual Inconsistencies in Attributable Text Generation [74.11403803488643]
  We introduce QASemConsistency, a new formalism for localizing factual inconsistencies in attributable text generation.<n>We show that QASemConsistency yields factual consistency scores that correlate well with human judgments.
  arXiv  Detail & Related papers  (2024-10-09T22:53:48Z)
• PARAPHRASUS : A Comprehensive Benchmark for Evaluating Paraphrase Detection Models [5.980822697955566]
  PARAPHR is a benchmark for multi-dimensional assessment, benchmarking and selection of paraphrase detection models.<n>We find that paraphrase detection models under our fine-grained evaluation lens exhibit trade-offs that cannot be captured through a single classification dataset.
  arXiv  Detail & Related papers  (2024-09-18T15:33:48Z)
• How to Prune Your Language Model: Recovering Accuracy on the "Sparsity May Cry'' Benchmark [60.72725673114168]
  We revisit the question of accurate BERT-pruning during fine-tuning on downstream datasets. We propose a set of general guidelines for successful pruning, even on the challenging SMC benchmark.
  arXiv  Detail & Related papers  (2023-12-21T03:11:30Z)
• Semi-DETR: Semi-Supervised Object Detection with Detection Transformers [105.45018934087076]
  We analyze the DETR-based framework on semi-supervised object detection (SSOD) We present Semi-DETR, the first transformer-based end-to-end semi-supervised object detector. Our method outperforms all state-of-the-art methods by clear margins.
  arXiv  Detail & Related papers  (2023-07-16T16:32:14Z)
• Beyond mAP: Towards better evaluation of instance segmentation [23.562251593257674]
  Average Precision does not penalize duplicate predictions in the high-recall range. We propose two new measures to explicitly measure the amount of both spatial and categorical duplicate predictions. Our Semantic Sorting and NMS can be added as a plug-and-play module to mitigate hedged predictions and preserve AP.
  arXiv  Detail & Related papers  (2022-07-04T17:56:14Z)
• MASKER: Masked Keyword Regularization for Reliable Text Classification [73.90326322794803]
  We propose a fine-tuning method, coined masked keyword regularization (MASKER), that facilitates context-based prediction. MASKER regularizes the model to reconstruct the keywords from the rest of the words and make low-confidence predictions without enough context. We demonstrate that MASKER improves OOD detection and cross-domain generalization without degrading classification accuracy.
  arXiv  Detail & Related papers  (2020-12-17T04:54:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.