.</p>\n","updatedAt":"2026-08-14T15:16:27.503Z","author":{"_id":"66350219843f549fdac86347","avatarUrl":"/avatars/1ad7aa4e8e6d80c5d50bf4de502f11a4.svg","fullname":"Matteo Negri","name":"MNegri","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":4,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"fr","probability":0.32275810837745667},"editors":["MNegri"],"editorAvatarUrls":["/avatars/1ad7aa4e8e6d80c5d50bf4de502f11a4.svg"],"reactions":[],"isReport":false}},{"id":"6a7fc2658df58acee67426c2","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":378,"isUserFollowing":false},"createdAt":"2026-08-15T01:35:33.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [Never Stop Speaking: a Denial-of-Service Attack on End-to-End Speech Language Models](https://huggingface.co/papers/2608.10405) (2026)\n* [From Semantics to Readout: Mechanistic Understanding of Audio Tokens after Fine-Tuning for Temporal Audio Grounding](https://huggingface.co/papers/2607.25355) (2026)\n* [Prosody-driven Jailbreaks in Audio LLMs: A Controlled Study and Mechanistic Analysis](https://huggingface.co/papers/2607.26541) (2026)\n* [Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning](https://huggingface.co/papers/2606.26036) (2026)\n* [Your Agentic LLMs Secretly Encode Latent Signals of Indirect Prompt-Injection Exposure](https://huggingface.co/papers/2608.02657) (2026)\n* [ReACT-CLIP: Response-Aware Test-Time Defense for Vision--Language Models](https://huggingface.co/papers/2608.01067) (2026)\n* [Model Confidence Under Answer-Preserving Attacks: An Informativeness-Manipulability Frontier](https://huggingface.co/papers/2608.06571) (2026)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"<p>This is an automated message from the <a href=\"https://huggingface.co/librarian-bots\">Librarian Bot</a>. I found the following papers similar to this paper. </p>\n<p>The following papers were recommended by the Semantic Scholar API </p>\n<ul>\n<li><a href=\"https://huggingface.co/papers/2608.10405\">Never Stop Speaking: a Denial-of-Service Attack on End-to-End Speech Language Models</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.25355\">From Semantics to Readout: Mechanistic Understanding of Audio Tokens after Fine-Tuning for Temporal Audio Grounding</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.26541\">Prosody-driven Jailbreaks in Audio LLMs: A Controlled Study and Mechanistic Analysis</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2606.26036\">Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2608.02657\">Your Agentic LLMs Secretly Encode Latent Signals of Indirect Prompt-Injection Exposure</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2608.01067\">ReACT-CLIP: Response-Aware Test-Time Defense for Vision--Language Models</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2608.06571\">Model Confidence Under Answer-Preserving Attacks: An Informativeness-Manipulability Frontier</a> (2026)</li>\n</ul>\n<p> Please give a thumbs up to this comment if you found it helpful!</p>\n<p> If you want recommendations for any Paper on Hugging Face checkout <a href=\"https://huggingface.co/spaces/librarian-bots/recommend_similar_papers\">this</a> Space</p>\n<p> You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: <code>@librarian-bot recommend</code></p>\n","updatedAt":"2026-08-15T01:35:33.071Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":378,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7014416456222534},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.09158","authors":[{"_id":"6a7f3135f747ea94019af553","name":"Yuanhe Zhang","hidden":false},{"_id":"6a7f3135f747ea94019af554","name":"Weiliu Wang","hidden":false},{"_id":"6a7f3135f747ea94019af555","name":"Jie Ren","hidden":false},{"_id":"6a7f3135f747ea94019af556","name":"Liang Lin","hidden":false},{"_id":"6a7f3135f747ea94019af557","name":"Zhenhong Zhou","hidden":false},{"_id":"6a7f3135f747ea94019af558","name":"Haoran Gao","hidden":false},{"_id":"6a7f3135f747ea94019af559","name":"Kun Wang","hidden":false},{"_id":"6a7f3135f747ea94019af55a","name":"Chen Li","hidden":false},{"_id":"6a7f3135f747ea94019af55b","name":"Li Sun","hidden":false},{"_id":"6a7f3135f747ea94019af55c","name":"Sen Su","hidden":false}],"publishedAt":"2026-08-10T00:00:00.000Z","submittedOnDailyAt":"2026-08-14T00:00:00.000Z","title":"From Inaudible Inputs to Model Failures: Low-Frequency Safety Risks in LALMs","submittedOnDailyBy":{"_id":"66350219843f549fdac86347","avatarUrl":"/avatars/1ad7aa4e8e6d80c5d50bf4de502f11a4.svg","isPro":false,"fullname":"Matteo Negri","user":"MNegri","type":"user","name":"MNegri"},"summary":"Large audio-language models (LALMs) have demonstrated strong capabilities in understanding diverse audio inputs. This diversity includes low-frequency signals that are inaudible to humans but can still enter the model and influence its generation. However, the practical impact of such low-frequency inputs on LALMs remains largely unexplored. In this paper, we propose Intermittent Low-Frequency Lockout (ILL), an inaudible red teaming method that evaluates this risk using a universal waveform template in a black box setting. ILL uses Sentence Attention Scale Estimation to determine active intervals and Frequency Confusion Transfer to construct a low-frequency waveform with continuous phase from corpus spectral variation. To mitigate this risk, we propose Distributional Requery Guard (DRG) to detect low-frequency distribution shifts and conditionally request a second recording for semantic recovery. Across six LALMs and multiple audio understanding tasks, ILL reduces accuracy by up to 67 percentage points while receiving a mean human audibility rating of 1.33, close to 1.17 for clean audio; DRG raises mean attacked accuracy from 28.5\\% to 46.1\\% after clean reacquisition. These findings identify a previously overlooked safety risk for LALMs and provide a foundation for future research on robust audio understanding.","upvotes":2,"discussionId":"6a7f3136f747ea94019af55d","ai_summary":"Researchers propose a black-box red-teaming method using inaudible low-frequency waveforms to expose vulnerabilities in audio-language models, alongside a defense that detects distribution shifts and requests a second recording to recover accuracy.","ai_keywords":["Large audio-language models","Intermittent Low-Frequency Lockout","Sentence Attention Scale Estimation","Frequency Confusion Transfer","Distributional Requery Guard","low-frequency red teaming","black-box attack"],"ai_summary_model":"thinkingmachines/Inkling-Small"},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"63ac5701c21e60a3e9b58aa7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ac5701c21e60a3e9b58aa7/g6EX7diOpuA94R2ab-rZC.png","isPro":true,"fullname":"Dipankar Sarkar","user":"dipankarsarkar","type":"user"},{"_id":"631e14ac473a6825f285e89d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/631e14ac473a6825f285e89d/K-6QnoeGLg8XFvbTMMdqA.jpeg","isPro":false,"fullname":"Yury Panikov","user":"panikov","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.09158.md","query":{}}">

hf papers read 2608.09158

curl -LsSf https://hf.co/cli/install.sh | bash