We build multi-agent safety & security environments for RL training and adaptive benchmarking.
Tübingen, Germany
Joined May 2026
- Tweets12
- Following5
- Followers478
- Likes12
Pinned Tweet
We're excited to announce our new work!
Asynchronous monitoring, incident investigations, and compliance audits primarily rely on agent traces to reconstruct a sequence of actions taken by agents. These analyses assume that LLM agents cannot tamper with their own execution traces. We show that local LLM agents such as Claude Code, Codex, Antigravity, Open Code and Grok Build fail to enforce this boundary. All tested harnesses, except Muse Code, allowed agents to delete their traces when asked, without triggering monitor guardrails.
This demonstrates a critical gap on two levels:
1. the low refusal rate on a model level suggests that current alignment training does not sufficiently discourage trace tampering, creating a risk of malicious or accidental manipulation of execution traces,
2. on the harness level, common agent CLIs and configurations also do not provide very effective safeguards against trace tampering.
At Exponential, we are actively working on security and safety benchmarks, both public and private. Get in touch if you are interested!
More details: expsec.ai.
We are proud to share that our two new benchmarks, RepoInject v1 and BrowserInject v1, were used by Meta TBD Labs to report the robustness of Muse Spark 1.3 to prompt injections.
The robustness of Muse Spark 1.3 is comparable to that of GPT-5.6 Sol—congrats to the team!
1/4
Results: On RepoInject v1, Muse Spark 1.3 and Claude Opus 5 show similar results, while GPT-5.6 Sol performs slightly better.
Muse Spark 1.3 generally blocks direct harmful commands, but more complex attacks still succeed while the model is performing normal coding tasks. This failure mode is also found in other models.
On BrowserInject v1, while Muse Spark 1.3 performs similarly to GPT-5.6 Sol and much better than Kimi K3, trailing Claude Opus 5.
3/4
We are working with frontier labs on security and safety evals that leverage adaptive attacks. Get in touch if you are interested!
More details about Exponential: expsec.ai/
Muse Spark 1.3 Evaluation Report: research.meta.ai/static/muse…
4/4
We're excited to announce a new paper!
We find that many recent models end up evading monitors under ordinary task pressure. We introduce EvasionBench to measure this behavior.
We find a very large variance in how models behave: from 88% evasion success of GLM-5.2 to 0% of GPT-6 Astra under best-of-3 evaluation. At the same time, GPT-6 Sol is at 20%.
We are actively working on security and safety benchmarks, both public and private. Get in touch if you are interested!
More details: expsec.ai/
In our first research release at Exponential Security Labs, we built an automated, multi-turn and multi-lingual red-teaming system and tested recent open-weight and proprietary models on purpose-built regional jailbreak datasets.🧵
Check out the results on our Sovereign-Jbreak dataset in our blogpost at expsec.ai/blog/automated-red…
The dataset can be found at huggingface.co/datasets/ExpS…
We are also interested in working with teams building models, by providing security-focused training data and RL environments for post-training. If this sounds interesting, please reach out to us at [email protected]
Today we are announcing our new startup: Exponential Security Labs.
AI agents are being deployed everywhere, making high-stakes decisions and increasingly automating research itself. Yet their reliability and security remain unsolved technical problems, on a frontier that keeps shifting. Our mission is to secure agentic systems through self-improving red-teaming and guardrail agents, leveraging autonomous AI research.
We sit at the intersection of safety and self-improvement, building on a decade of research in adversarial robustness and AI safety. We're looking for exceptional people to join us!
We're looking for exceptional people to join us on this mission, and we're also eager to talk to companies deploying AI agents who want to improve their security. Please get in touch!
More details: expsec.ai/
Founding team: @Nmndsingh, @AnselmPaulus, @maksym_andr, Matthias Hein
