{"total":29,"items":[{"citing_arxiv_id":"2605.23190","ref_index":9,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Hidden Human-Like Nature of Machine-Generated Texts: Theory and Detection Enhancement","primary_cat":"cs.CL","submitted_at":"2026-05-22T03:17:21+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Reveals hidden human-like spans in machine-generated texts that raise detection complexity and proposes a stacked enhancement framework with hard-EM optimization to improve detectors across LLMs.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.22654","ref_index":110,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Seeing the Poem: Image-Semantic Detection of AI-Generated Modern Chinese Poetry with MLLMs","primary_cat":"cs.CL","submitted_at":"2026-05-21T15:57:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"An image-semantic guided method enhances MLLMs for detecting AI-generated modern Chinese poetry by combining poem text with visual representations of content, achieving 85.65% Macro-F1 with Gemini and outperforming text baselines and RoBERTa.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.19722","ref_index":52,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Measuring Safety Alignment Effects in Autonomous Security Agents","primary_cat":"cs.CR","submitted_at":"2026-05-19T11:55:54+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A trace-based benchmark of 30 security tasks finds that less-restricted LLM derivatives outperform stock safety-aligned models on some agent tasks for Gemma but not Qwen or Llama, with similar patterns on non-security controls.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.16107","ref_index":7,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Multi-Level Contextual Token Relation Modeling for Machine-Generated Text Detection","primary_cat":"cs.CL","submitted_at":"2026-05-15T15:55:58+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A multi-level framework that models local and global relations among token detection scores to improve machine-generated text detection with low overhead.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.15518","ref_index":2,"ref_count":2,"confidence":0.98,"is_internal_anchor":true,"paper_title":"DetectRL-X: Towards Reliable Multilingual and Real-World LLM-Generated Text Detection","primary_cat":"cs.CL","submitted_at":"2026-05-15T01:29:26+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"DetectRL-X is a multilingual benchmark evaluating LLM text detectors on 8 languages, 6 domains, 4 commercial generators, and paraphrasing/perturbation attacks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.06903","ref_index":30,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"MELD: Multi-Task Equilibrated Learning Detector for AI-Generated Text","primary_cat":"cs.CL","submitted_at":"2026-05-07T20:05:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MELD is a multi-task AI-text detector using auxiliary heads, uncertainty-weighted losses, EMA distillation, and pairwise ranking that reaches 99.9% TPR at 1% FPR on a new held-out benchmark while remaining competitive on the RAID leaderboard.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"detectors are easy to deploy because they do not require detector-specific training, but their behavior is tied to the coverage and calibration of the reference models, making them sensitive to paraphrase and surface perturbations [9, 20]. Supervised encoder detectors.Supervised methods train discriminative models from labeled human and AI text. Early studies fine-tuned RoBERTa-style encoders [ 30]. Subsequent work improved this recipe with structured features [36], adversarial paraphrasing [17], stronger encoder backbones [8, 38], representation-based detection [7], and one-class objectives [43]. While these methods can perform well on in-distribution benchmarks, they are typically trained with a single binary head. This gives the encoder limited incentive to preserve generator, attack, or domain"},{"citing_arxiv_id":"2605.05950","ref_index":37,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Lightweight Stylistic Consistency Profiling: Robust Detection of LLM-Generated Textual Content for Multimedia Moderation","primary_cat":"cs.CL","submitted_at":"2026-05-07T09:58:00+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"LiSCP detects LLM-generated text via stylistic consistency profiling across paraphrased variants and reports up to 11.79% better cross-domain accuracy plus robustness to adversarial attacks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.03723","ref_index":12,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Segmenting Human-LLM Co-authored Text via Change Point Detection","primary_cat":"cs.CL","submitted_at":"2026-05-05T13:08:55+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Adapts change point detection to segment human-LLM co-authored text using weighted and generalized algorithms with minimax optimality and strong empirical results against baselines.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.00348","ref_index":26,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Block-wise Codeword Embedding for Reliable Multi-bit Text Watermarking","primary_cat":"cs.CR","submitted_at":"2026-05-01T02:14:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"BREW achieves TPR of 0.965 and FPR of 0.02 under 10% synonym substitution by shifting from ECC decoding to designated verification with block voting and local validation.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.26328","ref_index":22,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"DSIPA: Detecting LLM-Generated Texts via Sentiment-Invariant Patterns Divergence Analysis","primary_cat":"cs.CL","submitted_at":"2026-04-29T06:22:08+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"DSIPA is a zero-shot black-box detector that uses sentiment distribution consistency and preservation metrics to identify LLM text, reporting up to 49.89% F1 gains over baselines across domains and models.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.25860","ref_index":27,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Luminol-AIDetect: Fast Zero-shot Machine-Generated Text Detection based on Perplexity under Text Shuffling","primary_cat":"cs.CL","submitted_at":"2026-04-28T16:58:55+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.21223","ref_index":11,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Zero-Shot Detection of LLM-Generated Text via Implicit Reward Model","primary_cat":"cs.CL","submitted_at":"2026-04-23T02:37:34+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"IRM derives implicit reward signals from off-the-shelf LLMs to detect generated text zero-shot and reports better results than prior zero-shot and supervised detectors on the DetectRL benchmark.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.20131","ref_index":207,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Whose Story Gets Told? Positionality and Bias in LLM Summaries of Life Narratives","primary_cat":"cs.CL","submitted_at":"2026-04-22T02:58:51+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A proposed pipeline shows LLMs introduce detectable race and gender biases when summarizing life narratives, creating potential for representational harm in research.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16923","ref_index":10,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Alignment Imprint: Zero-Shot AI-Generated Text Detection via Provable Preference Discrepancy","primary_cat":"cs.AI","submitted_at":"2026-04-18T09:12:24+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"LAPD, derived from the provable preference discrepancy in aligned LLMs, improves zero-shot AI text detection by 45.82% over baselines with claimed statistical dominance over Fast-DetectGPT.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16058","ref_index":22,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"LLMSniffer: Detecting LLM-Generated Code via GraphCodeBERT and Supervised Contrastive Learning","primary_cat":"cs.SE","submitted_at":"2026-04-17T13:32:25+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"LLMSniffer improves detection of LLM-generated code on GPTSniffer and Whodunit benchmarks by fine-tuning GraphCodeBERT via two-stage supervised contrastive learning plus preprocessing and MLP classification.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.14111","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Interpretable Stylistic Variation in Human and LLM Writing Across Genres, Models, and Decoding Strategies","primary_cat":"cs.CL","submitted_at":"2026-04-15T17:31:21+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Genre and model exert stronger influence on writing style than human/LLM source or decoding strategy in a broad comparison of lexicogrammatical features.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.08888","ref_index":98,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"From OSS to Open Source AI: an Exploratory Study of Collaborative Development Paradigm Divergence","primary_cat":"cs.SE","submitted_at":"2026-04-10T02:52:09+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Open source AI shows lower collaboration intensity, reduced direct contributions, and a shift toward adaptive use rather than joint improvement compared to traditional OSS.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"In recent years, Artificial Intelligence has experienced unprecedented advancement, drawing sub- stantial attention and investment from researchers and commercial enterprises. The development of AI models is increasingly embracing open-source paradigm: companies engaged in AI development increasingly demonstrate a preference for open-sourcing their models [98, 111], and concurrently, numerous smaller collectives and individual developers have shared their models with the global community, establishing this as a defining trend of the current technological landscape [59, 81]. Here, we use the term Open Source AI Model (OSM) to refer to AI model whose components, ∗Corresponding author. Authors' Contact Information: Hengzhi Ye, hzye@stu."},{"citing_arxiv_id":"2601.22002","ref_index":47,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Rate-Distortion Optimization for Transformer Inference","primary_cat":"cs.LG","submitted_at":"2026-01-29T17:12:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"A rate-distortion framework for lossy compression of transformer representations yields substantial bitrate savings on language tasks while preserving accuracy, with observed rates aligning to derived information-theoretic bounds.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2511.12468","ref_index":9,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Detecting LLM-Assisted Academic Dishonesty using Keystroke Dynamics","primary_cat":"cs.HC","submitted_at":"2025-11-16T05:53:03+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Keystroke dynamics models outperform text-only detectors for spotting LLM-assisted academic dishonesty in practical scenarios, though performance drops under adversarial conditions.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2511.05501","ref_index":62,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Towards Real-World Validity in Generative AI Benchmarks: Understanding and Designing Domain-Centered Evaluations for Journalism Practitioners","primary_cat":"cs.HC","submitted_at":"2025-09-30T21:36:23+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A human-centered design workshop with journalism practitioners yields an evaluation cookbook and design requirements for contextualized, value-aligned generative AI benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2502.11336","ref_index":32,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"ExaGPT: Example-Based Machine-Generated Text Detection for Human Interpretability","primary_cat":"cs.CL","submitted_at":"2025-02-17T01:15:07+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"ExaGPT uses span-level similarity retrieval from human and LLM datastores to detect machine-generated text while supplying the matching spans as human-interpretable evidence, achieving up to 37-point accuracy gains over prior interpretable detectors at 1% FPR.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2410.23728","ref_index":66,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"GigaCheck: Detecting LLM-generated Content via Object-Centric Span Localization","primary_cat":"cs.CL","submitted_at":"2024-10-31T08:30:55+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"GigaCheck detects LLM-generated text at both document and span levels by combining fine-tuned language-model embeddings with a DETR-like architecture that treats generated intervals as detectable objects.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2406.10162","ref_index":300,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models","primary_cat":"cs.AI","submitted_at":"2024-06-14T16:26:20+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"LLMs trained on simple specification gaming generalize to zero-shot reward tampering including rewriting their own reward function.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2306.12001","ref_index":106,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"An Overview of Catastrophic AI Risks","primary_cat":"cs.CY","submitted_at":"2023-06-21T03:35:06+00:00","verdict":"ACCEPT","verdict_confidence":"MODERATE","novelty_score":3.0,"formal_verification":"none","one_line_summary":"The paper categorizes sources of catastrophic AI risks into malicious use, AI race, organizational risks, and rogue AIs, providing illustrative stories and mitigation suggestions for each.","context_count":1,"top_context_role":"background","top_context_polarity":"support","context_text":"AI labs should acquire information about the safety of AI systems before making them available for broader use. One way to do this is to commission red teams to find hazards before AI systems are promoted to production. AI labs can execute a \"staged release\": gradually expanding access to the AI system so that safety failures are fixed before they produce widespread negative consequences [106]. Finally, AI labs can avoid deploying or training more powerful AI systems until currently deployed AI systems have proven to be safe over time. Publication reviews. AI labs have access to potentially dangerous or dual-use information such as model weights and research intellectual property (IP) that would be dangerous if proliferated. An internal review"},{"citing_arxiv_id":"2303.11156","ref_index":88,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Can AI-Generated Text be Reliably Detected?","primary_cat":"cs.CL","submitted_at":"2023-03-17T17:53:19+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Recursive paraphrasing attacks substantially lower detection rates for multiple AI text detectors with only minor quality loss, while a theoretical analysis ties best-case AUROC to total variation distance between human and AI distributions.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2112.04359","ref_index":260,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Ethical and social risks of harm from Language Models","primary_cat":"cs.CL","submitted_at":"2021-12-08T16:09:48+00:00","verdict":"ACCEPT","verdict_confidence":"MODERATE","novelty_score":6.0,"formal_verification":"none","one_line_summary":"The authors provide a detailed taxonomy of 21 risks associated with language models, covering discrimination, information leaks, misinformation, malicious applications, interaction harms, and societal impacts like job loss and environmental costs.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2110.08207","ref_index":60,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Multitask Prompted Training Enables Zero-Shot Task Generalization","primary_cat":"cs.LG","submitted_at":"2021-10-15T17:08:57+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Multitask fine-tuning of an encoder-decoder model on prompted datasets produces zero-shot generalization that often beats models up to 16 times larger on standard benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2102.04664","ref_index":73,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generation","primary_cat":"cs.SE","submitted_at":"2021-02-09T06:16:25+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"CodeXGLUE supplies a standardized collection of 10 code-related tasks, 14 datasets, an evaluation platform, and BERT-, GPT-, and encoder-decoder-style baselines.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2005.11401","ref_index":57,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks","primary_cat":"cs.CL","submitted_at":"2020-05-22T21:34:34+00:00","verdict":"ACCEPT","verdict_confidence":"MODERATE","novelty_score":7.0,"formal_verification":"none","one_line_summary":"RAG models set new state-of-the-art results on open-domain QA by retrieving Wikipedia passages and conditioning a generative model on them, while also producing more factual text than parametric baselines.","context_count":1,"top_context_role":"baseline","top_context_polarity":"baseline","context_text":"retrieved evidence. In many real-world applications, retrieval supervision signals aren't available, and models that do not require such supervision will be applicable to a wider range of tasks. We explore two variants: the standard 3-way classification task (supports/refutes/not enough info) and the 2-way (supports/refutes) task studied in Thorne and Vlachos [57]. In both cases we report label accuracy. 4 Results 4.1 Open-domain Question Answering Table 1 shows results for RAG along with state-of-the-art models. On all four open-domain QA tasks, RAG sets a new state of the art (only on the T5-comparable split for TQA). RAG combines the generation flexibility of the \"closed-book\" (parametric only) approaches and the performance of"}],"limit":50,"offset":0}