CS PhD student @TelAvivUni. Intern @google. Interested in buzzwords like AI and Security, and wherever they meet.
Joined October 2009
- Tweets113
- Following1.3K
- Followers137
- Likes7.2K
Pinned Tweet
Discrete trigger optimizers (eg GCG) are extremely useful in NLP security (eg LLM jailbreaks), but in practice it's an engineering mess.
Introducing TROPT: a powerful framework housing 15+ popular optimizers, ready to use against any objective (!) and any model (!!)
🧵
We found Claude Code's safety classifier (Auto Mode) can be nullified by a short, universal prompt injection in a repo's AGENTS.md -- making it approve unsafe (`rm -rf /`) and user-forbidden commands.
🧵
Overall, this attack highlights:
(1) the risk of untrusted sources entering the context of safety-critical LLMs;
(2) the importance of adaptive security evaluations;
(3) the brittleness of LLM-based classifiers.
Regardless of guardrails, sandbox your agents 🤖
Check out the blog post for more details: matanbt.github.io/post/yesmo…
Excited to share that TROPT has been accepted to NeurIPS 2026 (E&D Track)! 🥳
If you're using discrete text optimization methods for jailbreak / other tasks, check it out 👇
Matan Ben-Tov retweeted
We can finally talk about it:
We found a way to extract hidden reasoning of frontier models using a vulnerability in the APIs of every frontier AI company.
We verified that our reasoning token count matches billed API thinking tokens 1:1 for most of the prompts we queried.
I'll be in San Diego for #ACL2026 🇺🇸, presenting our TACL work on LLM jailbreaks through mech interp lens: Universal Jailbreak Suffixes Are Strong Attention Hijackers
@mahmoods01 @megamor2
📍Jul 5 14:00-14:10, Harbor D (Oral Session)
If you're interested in NLP security, mech interp, or leveraging discrete input optimizers for various uses, let's chat! 🌊
Matan Ben-Tov retweeted
Here we go! ✈️ 🇺🇸🇰🇷 If you're attending #ACL2026 and/or #ICML2026 check out recent work from our group with collaborators:
📍Jul 5 14:00-15:30: Matan will present his TACL work "Universal Jailbreak Suffixes Are Strong Attention Hijackers" @matanbt @mahmoods01
arxiv.org/abs/2506.12880
📍Jul 5 16:00-17:30: My student Or will show how matrix factorization can reveal compositional neuron groups and hierarchies in MLPs. @OrShafran
arxiv.org/abs/2506.10920
📍Jul 6 9:00-10:30: Tomer will talk about "MoNaCo: More Natural and Complex Questions for Reasoning Across Dozens of Documents" @TomerWolfson
arxiv.org/abs/2508.11133
📍Jul 7 14:00-15:45: Hadas will present our position: "Interpretability Can Be Actionable" @OrgadHadas
arxiv.org/abs/2605.11161
(will also appear at the MI workshop)
📍Jul 8 10:30-12:15: I will be presenting a position work with @_galyo and @ymatias on metacognition as a way to move beyond hallucinations
arxiv.org/abs/2605.01428
📍Jul 9 17:00-18:45: Or (after a transpacific flight from ACL) will present MFA: an unsupervised approach to disentangle representations of LLMs at scale using local geometry @OrShafran
arxiv.org/abs/2602.02464
(will also appear at the MI workshop)
Check out our paper & code:
🧑💻 pip install tropt
📄Paper: arxiv.org/abs/2606.23496
💻Code: github.com/matanbt/TROPT/
🌐 Website: tropt.dev
Happy TROPTing.
Huge thanks to @mahmoods01 for guiding me on this project.
I've long worked on this repo, and am very excited to see how it’ll be used!
I intend to keep growing it and iterate on it in light of any future community feedback -- do feel free to reach out and contribute!