SinoGlyphBench: A Diagnostic Benchmark for Chinese Glyph-Level Obfuscation in Language-Model Moderation

• URL: http://arxiv.org/abs/2609.05843v1
• Date: Sat, 05 Sep 2026 03:15:49 GMT
• Title: SinoGlyphBench: A Diagnostic Benchmark for Chinese Glyph-Level Obfuscation in Language-Model Moderation
• Abstract summary: Glyph-level obfuscation can leave harmful Chinese content readable to humans while degrading automated moderation.<n>We introduce SinoGlyphBench, a diagnostic benchmark that identifies label-critical semantic anchors and creates matched original and glyph-obfuscated inputs.
• Score: 25.91174070621304
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Glyph-level obfuscation can leave harmful Chinese content readable to humans while degrading automated moderation. We introduce SinoGlyphBench, a diagnostic benchmark that identifies label-critical semantic anchors and creates matched original and glyph-obfuscated inputs in text and image modalities. By perturbing anchors, background context, or both, this design distinguishes corruption of moderation-relevant evidence from general surface variation. Across 176,916 paired evaluations of 12 LLMs and MLLMs, obfuscation increases harmful false-negative and false-positive rates by 6.1 and 4.7 percentage points, respectively, and reduces four-way accuracy by 5.0 points. Models retain 75.7% of the decisions that were correct on the matched original inputs. Full-scope perturbations cause the largest degradation, anchor-only perturbations are more damaging than background-only perturbations, and cross-script substitution is particularly difficult in the text modality. Analysis of structured outputs identifies observable mismatches in visible-form reading, intended-message recovery, and final safety judgment. The evaluated models, therefore, remain brittle to Chinese content written with non-canonical glyphs. Resources are available at https://github.com/fengshun124/SinoGlyphBench.

Related papers

• Quantifying and Mitigating Korean Jamo-Level Typographical Vulnerabilities in Large Language Models [12.893256474784499]
  Korean introduces an additional typographical perturbation level not captured by ordinary character-level edit models.<n>Because syllable blocks are internally composed of sub-character units called jamo, keyboard-level errors can occur within a syllable, either producing a valid but semantically altered character or exposing raw jamo on the surface.<n>To quantify this vulnerability, we apply five jamo-level perturbation types to the KMMLU benchmark and evaluate four language models.<n>We propose Typo-Aware Chain-of-Thought (TACoT), which routes inputs to chain-of-thought inference only when the probe detects a likely typo
  arXiv  Detail & Related papers  (2026-08-31T04:28:14Z)
• Stoicheia: Character-Level Masked Diffusion for Ancient Greek Textual Restoration, Parsing, and Metrical Scansion [0.0]
  Stoicheia is a character-level masked-diffusion encoder for Ancient Greek.<n>We pretrain it on an open, revision-pinned corpus of 380M words and release eleven checkpoints.<n>On Ithaca's own test split, with identical frozen samples and strict scoring, Stoicheia reduces character error relative to both prior state-of-the-art systems.
  arXiv  Detail & Related papers  (2026-08-07T14:07:43Z)
• Aligned in Form, Not in Meaning: The Comprehension - Containment Decoupling of LLM Safety in Low-Resource Bangla Derogatory Speech [0.0]
  We audit five large language models on native Bangla derogatory speech (gali) across six protocols.<n>We propose that contemporary safety alignment is bound to high-resource surface forms rather than harmful meaning.
  arXiv  Detail & Related papers  (2026-08-03T23:09:19Z)
• CORE: Context-Robust Remasking for Diffusion Language Models [51.59514489363897]
  We propose Context-Robust Remasking (CORE), a training-free framework for inference-time revision.<n>Rather than trusting static token probabilities, CORE identifies context-brittle tokens by probing their sensitivity to targeted masked-context perturbations.<n>On LLaDA-8B-Base, CORE delivers consistent improvements across reasoning and code benchmarks, outperforming compute-matched baselines and improving MBPP by up to 9.2 percentage points.
  arXiv  Detail & Related papers  (2026-02-04T00:12:30Z)
• MirrorMark: A Distortion-Free Multi-Bit Watermark for Large Language Models [5.735801967350819]
  We propose MirrorMark, a distortion-free watermark for large language models (LLMs)<n>MirrorMark embeds multi-bit messages without altering the token probability distribution, preserving text quality by design.<n> Experiments show that MirrorMark matches the text quality of non-watermarked generation while achieving substantially stronger detectability.
  arXiv  Detail & Related papers  (2026-01-29T19:10:48Z)
• CLUE: Non-parametric Verification from Experience via Hidden-State Clustering [64.50919789875233]
  We show that correctness of a solution is encoded as a geometrically separable signature within the trajectory of hidden activations.<n>ClUE consistently outperforms LLM-as-a-judge baselines and matches or exceeds modern confidence-based methods in reranking candidates.
  arXiv  Detail & Related papers  (2025-10-02T02:14:33Z)
• Languages in Multilingual Speech Foundation Models Align Both Phonetically and Semantically [58.019484208091534]
  Cross-lingual alignment in pretrained language models (LMs) has enabled efficient transfer in text-based LMs.<n>It remains an open question whether findings and methods from text-based cross-lingual alignment apply to speech.
  arXiv  Detail & Related papers  (2025-05-26T07:21:20Z)
• Group-Adaptive Threshold Optimization for Robust AI-Generated Text Detection [58.419940585826744]
  We introduce FairOPT, an algorithm for group-specific threshold optimization for probabilistic AI-text detectors.<n>We partitioned data into subgroups based on attributes (e.g., text length and writing style) and implemented FairOPT to learn decision thresholds for each group to reduce discrepancy.<n>Our framework paves the way for more robust classification in AI-generated content detection via post-processing.
  arXiv  Detail & Related papers  (2025-02-06T21:58:48Z)
• Extract Free Dense Misalignment from CLIP [7.0247398611254175]
  This work proposes a novel approach, dubbed CLIP4DM, for detecting dense misalignments from pre-trained CLIP.<n>We revamp the gradient-based attribution computation method, enabling negative gradient of individual text tokens to indicate misalignment.<n>Our method demonstrates state-of-the-art performance among zero-shot models and competitive performance with fine-tuned models.
  arXiv  Detail & Related papers  (2024-12-24T12:51:05Z)
• Synchronous Faithfulness Monitoring for Trustworthy Retrieval-Augmented Generation [96.78845113346809]
  Retrieval-augmented language models (RALMs) have shown strong performance and wide applicability in knowledge-intensive tasks. This paper proposes SynCheck, a lightweight monitor that leverages fine-grained decoding dynamics to detect unfaithful sentences. We also introduce FOD, a faithfulness-oriented decoding algorithm guided by beam search for long-form retrieval-augmented generation.
  arXiv  Detail & Related papers  (2024-06-19T16:42:57Z)
• Provably Secure Disambiguating Neural Linguistic Steganography [66.30965740387047]
  The segmentation ambiguity problem, which arises when using language models based on subwords, leads to occasional decoding failures.<n>We propose a novel secure disambiguation method named SyncPool, which effectively addresses the segmentation ambiguity problem.<n> SyncPool does not change the size of the candidate pool or the distribution of tokens and thus is applicable to provably secure language steganography methods.
  arXiv  Detail & Related papers  (2024-03-26T09:25:57Z)

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.