TROPT is a Textual Trigger Optimization Toolbox for executing and developing discrete text optimizers that elicit (un)desired behaviors for various types of NLP models (LLMs, embeddings, classifiers) and applications (red-teaming, interpretability, etc.).</p>\n<p>Use it! :<br>🧑💻 <a href=\"https://github.com/matanbt/TROPT\" rel=\"nofollow\">https://github.com/matanbt/TROPT</a></p>\n","updatedAt":"2026-06-23T21:16:55.907Z","author":{"_id":"635671cdef1d4c919152b8e8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/635671cdef1d4c919152b8e8/kXi7uO9z_Et5vbJQTWfW5.jpeg","fullname":"Matan BT","name":"MatanBT","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":1,"identifiedLanguage":{"language":"en","probability":0.8513733148574829},"editors":["MatanBT"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/635671cdef1d4c919152b8e8/kXi7uO9z_Et5vbJQTWfW5.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.23496","authors":[{"_id":"6a3af5700a86ac3098d5d5c6","name":"Matan Ben-Tov","hidden":false},{"_id":"6a3af5700a86ac3098d5d5c7","name":"Mahmood Sharif","hidden":false}],"publishedAt":"2026-06-22T00:00:00.000Z","submittedOnDailyAt":"2026-06-23T00:00:00.000Z","title":"TROPT: An Open Framework for Unifying and Advancing Discrete Text Optimization","submittedOnDailyBy":{"_id":"635671cdef1d4c919152b8e8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/635671cdef1d4c919152b8e8/kXi7uO9z_Et5vbJQTWfW5.jpeg","isPro":false,"fullname":"Matan BT","user":"MatanBT","type":"user","name":"MatanBT"},"summary":"Discrete text-trigger optimization -- searching for text sequences that, when ingested by a model, steer it toward a specified objective -- underpins model red-teaming (e.g., LLM jailbreaks), as well as auditing and interpretability. However, the current state of discrete optimizers hinders their adoption and progress. First, existing optimizers, when open-sourced at all, are scattered across research codebases tied to specific models, objectives, and problem domains. Second, optimizer variants proliferate, each requiring engineering overhead to use or extend, and remaining hard to compare head-to-head. Together, these raise the bar for adopting optimizers in existing or new domains, and for advancing them via new strategies. We address these gaps with TROPT, the first open-source framework that unifies discrete optimizers' execution and standardizes their development under a single interface. TROPT makes it easy to customize end-to-end optimization recipes by swapping any component -- models, objectives, and optimizers -- extending its reach across domains and new applications. TROPT currently ships with 30+ optimization recipes -- covering applications such as jailbreaking and probing model internals -- built from 15+ optimizers (spanning white-box to black-box access) and 15+ losses, from foundational to state-of-the-art methods. Demonstrating its utility, we leverage TROPT in several studies: (i) controlled, large-scale experiments comparing and enhancing optimization strategies for LLM jailbreaks, revealing potent-yet-underadopted techniques; and (ii) porting optimizers from one domain (e.g., LLM jailbreak) to new domains (e.g., corpus-poisoning embedding model). In all, TROPT significantly lowers the barrier to adopting and advancing discrete text optimization.","upvotes":1,"discussionId":"6a3af5700a86ac3098d5d5c8","projectPage":"https://tropt.dev","githubRepo":"https://github.com/matanbt/TROPT","githubRepoAddedBy":"user","ai_summary":"A unified open-source framework for discrete text-trigger optimization that standardizes the development and execution of optimization strategies across various domains and applications.","ai_keywords":["discrete text-trigger optimization","model red-teaming","LLM jailbreaks","auditing","interpretability","optimization recipes","white-box access","black-box access","loss functions","optimization strategies"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct","githubStars":5,"organization":{"_id":"6107dfc57602f8e9ed8bb5cb","name":"tau","fullname":"Tel Aviv University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1628143727824-610b729f9da682cd54ad9adf.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"635671cdef1d4c919152b8e8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/635671cdef1d4c919152b8e8/kXi7uO9z_Et5vbJQTWfW5.jpeg","isPro":false,"fullname":"Matan BT","user":"MatanBT","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6107dfc57602f8e9ed8bb5cb","name":"tau","fullname":"Tel Aviv University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1628143727824-610b729f9da682cd54ad9adf.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2606/2606.23496.md","query":{}}">
TROPT: An Open Framework for Unifying and Advancing Discrete Text Optimization
Abstract
A unified open-source framework for discrete text-trigger optimization that standardizes the development and execution of optimization strategies across various domains and applications.
Discrete text-trigger optimization -- searching for text sequences that, when ingested by a model, steer it toward a specified objective -- underpins model red-teaming (e.g., LLM jailbreaks), as well as auditing and interpretability. However, the current state of discrete optimizers hinders their adoption and progress. First, existing optimizers, when open-sourced at all, are scattered across research codebases tied to specific models, objectives, and problem domains. Second, optimizer variants proliferate, each requiring engineering overhead to use or extend, and remaining hard to compare head-to-head. Together, these raise the bar for adopting optimizers in existing or new domains, and for advancing them via new strategies. We address these gaps with TROPT, the first open-source framework that unifies discrete optimizers' execution and standardizes their development under a single interface. TROPT makes it easy to customize end-to-end optimization recipes by swapping any component -- models, objectives, and optimizers -- extending its reach across domains and new applications. TROPT currently ships with 30+ optimization recipes -- covering applications such as jailbreaking and probing model internals -- built from 15+ optimizers (spanning white-box to black-box access) and 15+ losses, from foundational to state-of-the-art methods. Demonstrating its utility, we leverage TROPT in several studies: (i) controlled, large-scale experiments comparing and enhancing optimization strategies for LLM jailbreaks, revealing potent-yet-underadopted techniques; and (ii) porting optimizers from one domain (e.g., LLM jailbreak) to new domains (e.g., corpus-poisoning embedding model). In all, TROPT significantly lowers the barrier to adopting and advancing discrete text optimization.
Community
TROPT is a Textual Trigger Optimization Toolbox for executing and developing discrete text optimizers that elicit (un)desired behaviors for various types of NLP models (LLMs, embeddings, classifiers) and applications (red-teaming, interpretability, etc.).
Use it! :
🧑💻 https://github.com/matanbt/TROPT
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2606.23496 in a model README.md to link it from this page.
Cite arxiv.org/abs/2606.23496 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2606.23496 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.