If you care about rigorous evaluation of agentic systems, give it a look at MASEval! The harness is an important element of agents. MASEval makes it straightforward to change its components and evaluate their impact. MASEval is our first software! parameterlab.github.io/MASEval/ ⬇️

C Emde@cemde.bsky.social · 7mo ago

1/ Evaluating a single agent harness is hard. Evaluating a multi-agent system? Whole different problem. Most eval tools treat the model as the unit of analysis. In multi-agent systems, the system is what matters. That's why we built MASEval 🧵 #AI #Agents #Eval #MultiAgentSystem #LLM

‼️New paper from Parameter Lab! ⛓️‍💥 We identify privacy collapse, a silent failure mode of LLMs: LLMs fine-tuned on seemingly benign data can lose their ability to respect contextual privacy norms. Done by @anmolgoel.bsky.social during his internship! Check-out 👇

Anmol Goel@anmolgoel.bsky.social · 9mo ago

🚨 Fine-tuning your model to be more helpful or empathetic might be making it less private, without you noticing. In our latest work, we show that benign fine-tuning can silently break contextual privacy in language models while safety & general capabilities appear intact. ⬇️

👏 Proud to share that the paper that Ahmed Heakl authored during his internship at Parameter Lab was accepted at #ICLR2026! See how 🩺Dr.LLM increases accuracy and decreases inference computations of frozen LLMs: www.linkedin.com/posts/ahmed-...

#llm #ai #efficientai #nlp #mlresearch #reasoning #adaptivecompute | Ahmed Heakl | 19 comments

Super excited to share the last work from my internship in Germany 🇩🇪! 🚀 Dr.LLM: Dynamic Layer Routing for LLMs > What if we can reduce computation AND increase accuracy? 🤯 Most prompts don’t need ev...

linkedin.com

Our #EMNLP2025 paper Leaky Thoughts 🫗 shows that Large Reasoning Models (LRMs) can unintentionally leak sensitive information hidden in their internal thoughts. 📍 Come chat with Tommaso at our poster on Friday 7th, 10:30–12:00 in Hall C3 📄 aclanthology.org/2025.emnlp-m...

Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers

Tommaso Green, Martin Gubri, Haritz Puerto, Sangdoo Yun, Seong Joon Oh. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

aclanthology.org

Parameter Lab@parameterlab.bsky.social · last yr.

🫗 An LLM's "private" reasoning may leak your sensitive data! 🎉 Excited to share our paper "Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers" was accepted at #EMNLP main! 1/2

🫗 An LLM's "private" reasoning may leak your sensitive data! 🎉 Excited to share our paper "Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers" was accepted at #EMNLP main! 1/2

🔎Does Conversational SEO actually work? Our new benchmark has an answer! Excited to announce our new paper: C-SEO Bench: Does Conversational SEO Work? 🌐 RTAI: researchtrend.ai/papers/2506.... 📄 Paper: arxiv.org/abs/2506.11097 💻 Code: github.com/parameterlab... 📊 Data: huggingface.co/datasets/par...

Excited to share that our paper "Scaling Up Membership Inference: When and How Attacks Succeed on LLMs" will be presented next week at #NAACL2025! 🖼️ Catch us at Poster Session 8 - APP: NLP Applications 🗓️ May 2, 11:00 AM - 12:30 PM 🗺️ Hall 3 Hope to see you there!

Martin Gubri@mgubri.bsky.social · 2y ago

📄 Excited to share our latest paper on the scale required for successful membership inference in LLMs! We investigate a continuum from single sentences to large document collections. Huge thanks to an incredible team: Haritz Puerto, @coallaoh.bsky.social and @oodgnas.bsky.social!

👥 We're Hiring: Senior/Junior Data Engineer! 📍 Remote or Local | Full-Time or Part-Time At ResearchTrend.AI, we’re building a platform that connects researchers and AI engineers worldwide—helping them stay ahead with daily digests, insightful summaries, and interactive events.

ResearchTrend.AI

Explore the most trending research topics in AI

researchtrend.ai

🔎 Wonder how to prove an LLM was trained on a specific text? The camera ready of our Findings of #NAACL 2025 paper is available! 📌 TLDR: longs texts are needed to gather enough evidence to determine whether specific data points were included in training of LLMs: arxiv.org/abs/2411.00154

Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models

Membership inference attacks (MIA) attempt to verify the membership of a given data sample in the training set for a model. MIA has become relevant in recent years, following the rapid development of ...

arxiv.org

Martin Gubri@mgubri.bsky.social · 2y ago

📄 Excited to share our latest paper on the scale required for successful membership inference in LLMs! We investigate a continuum from single sentences to large document collections. Huge thanks to an incredible team: Haritz Puerto, @coallaoh.bsky.social and @oodgnas.bsky.social!

We are delighted to announce that our research paper on the scale of LLM membership inference has been accepted for publication in the Findings of #NAACL2025! 🎉

Martin Gubri@mgubri.bsky.social · 2y ago

📄 Excited to share our latest paper on the scale required for successful membership inference in LLMs! We investigate a continuum from single sentences to large document collections. Huge thanks to an incredible team: Haritz Puerto, @coallaoh.bsky.social and @oodgnas.bsky.social!

There's an internship opening at @parameterlab.bsky.social : parameterlab.de/careers The research outputs have been quite successful so far: researchtrend.ai/organization...

Careers | Parameter Lab

Join us at Parameter Lab to shape the future of safe AI. In our dynamic and inclusive environment, we focus not only on our mission but also on fostering your personal growth through rewarding work ex...

parameterlab.de

🎉We’re pleased to share the release of the models from our Apricot🍑 paper, accepted at ACL 2024! At Parameter Lab, we believe openness and reproducibility are essential for advancing science, and we've put in our best effort to ensure it. 🤗 huggingface.co/collections/... 🧵 bsky.app/profile/dnns...

🍑 Apricot Models - a parameterlab Collection

Fine-tuned models for black-box LLM calibration, trained for "Apricot: Calibrating Large Language Models Using Their Generations Only" (ACL 2024)

huggingface.co

Dennis Ulmer @ACL26@dnnslmr.bsky.social · 3y ago

Obtaining calibrated confidence scores from LLMs is hard, especially for black-box models. So, can we maybe predict them directly from the generated text? 🤔 Internship work at Parameter Lab with Martin Gubri, Sangdoo Yun, Hwaran Lee, Seong Joon Oh! arxiv.org/abs/2403.059... [1/6]

🚨📄 Exciting new research! Discover when and at what scale we can detect if specific data was used in training LLMs — a method known as Membership Inference (MIA)! Our findings open new doors for using MIA as potential legal evidence in AI. 🧵 https://arxiv.org/abs/2411.00154

📄 We’ve been working on the calibration of confidence score for black-box LLMs. See the thread below for an overview of the 🍑 Apricot paper, proudly accepted at #ACL24! #uncertainty

Dennis Ulmer @ACL26@dnnslmr.bsky.social · 3y ago

Obtaining calibrated confidence scores from LLMs is hard, especially for black-box models. So, can we maybe predict them directly from the generated text? 🤔 Internship work at Parameter Lab with Martin Gubri, Sangdoo Yun, Hwaran Lee, Seong Joon Oh! arxiv.org/abs/2403.059... [1/6]

Check out one of our latest papers about LLM fingerprinting!

Martin Gubri@mgubri.bsky.social · 2y ago

🌟 Pleased to join Bluesky! As a first post, allow me to share my latest first-author paper, TRAP 🪤, presented at #ACL24 (findings). 🦹💥 We explore how to detect if an LLM was stolen or leaked🤖💥 We showcase how to use adversarial prompt as #fingerprint for #LLM. A thread 🧵 ⬇️⬇️⬇️

We are excited to join Bluesky! At Parameter Lab, we're committed to enhancing AI safety and trustworthiness. Our research addresses privacy, copyright, and security challenges in foundational models. Follow us for insights and updates on trustworthy AI research!