Parameter Lab

Empowering individuals and organisations to safely use foundational AI models.

[Parameter Lab @parameterlab]

‼️ 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 @anmgoel during his internship!

Check-out 👇

[Anmol Goel @anmgoel]

🚨 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.

⬇️

[Anmol Goel @anmgoel]

🚨 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.

⬇️

[Anmol Goel @anmgoel]

For more insights:
🌐 Project page: https://t.co/pf9sIF326u
📄 Paper: https://t.co/yq1t53luKl

Work done with the amazing team at @parameterlab: @CorEmde, @oodgnas, @coallaoh and @framart1 with support from @NAVER_AI_Lab.
#NLProc #AISafety #Privacy #LLMs

[Parameter Lab @parameterlab]

👏 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: https://t.co/wde7gVzVKc

[Martin Gubri @EMNLP @framart1]

🎉Delighted to announce that our 🫗Leaky Thoughts paper about contextual privacy with reasoning models is accepted to #EMNLP main!
Huge congrats to the amazing team @tommasogreen @HaritzPuerto @coallaoh @oodgnas

[Parameter Lab @parameterlab]

🧪 Our latest research: Does SEO boost the visibility of content in LLM-based conversational search?

We present C-SEO Bench, a benchmark to evaluate conversational SEO strategies.

Key takeaway: SEO methods that target LLM do not work. But surprisingly, traditional SEO is not dead: it still matters, as LLMs tend to favour content already ranked higher in their input.

[Haritz Puerto @HaritzPuerto]

🔎 Does Conversational SEO (C-SEO) actually work? Our new benchmark has an answer.
Excited to announce C-SEO Bench: Does Conversational SEO Work?
🌐 RTAI: https://t.co/cCxxtfxiM5
📄 Paper: https://t.co/rnOdAPhUQy
💻 Code: https://t.co/PyCGJAGU72
📊 Data: https://t.co/cICQlNCB5c

[Parameter Lab @parameterlab]

Authored by @tommasogreen @framart1 @HaritzPuerto @oodgnas @coallaoh
Huge thanks to this amazing team for the great work!

[Parameter Lab @parameterlab]

🎉 Very excited to see our new Leaky Thoughts 🫗 paper featured among last week's top AI papers by both @dair_ai and @TheAITimeline!

- https://t.co/t6xpZCaO92
- https://t.co/LgBk52ay76

➡️ Learn more about the paper in this great thread by @omarsar0: https://t.co/qNwUoLhGJ2
➡️ ArXiv link: https://t.co/v6KWArwOja

[elvis @omarsar0]

Leaky Thoughts

Hey AI devs, be careful how you prompt reasoning models.

This work shows that reasoning traces frequently contain sensitive user data.

More of my notes below:

[Haritz Puerto @HaritzPuerto]

We see news like this from time to time and that’s why it’s vital to keep researching on tools to prove these cases! Our #NAACL2025 paper shows that with over 10k tokens, we can reliably detect whether a text was part of an LLM’s training data https://t.co/G0iXafs4In

[Lucas Beyer (bl16) @giffmana]

Proud to have co-led more than 200 of those 999 during my time at Google.

[Haritz Puerto @HaritzPuerto]

Do you want to prove that your copyrighted document/corpus was trained by an LLM? Come to poster 46 #NAACL2025

[Haritz Puerto @HaritzPuerto]

#NAACL2025 has started! I’ll be presenting my work at @parameterlab about detecting pretraining data on Friday

🗓️ May 2, 11:00 AM - May 2, 12:30 PM
🗺️ Poster Session 8 - APP: NLP Applications
Location: Hall 3

Work with @framart1 @oodgnas @coallaoh

[Haritz Puerto @HaritzPuerto]

I will be in person at #NAACL2025 🌵🇺🇸 to present Scaling Up Membership Inference: When and How Attacks Succeed on LLMs. Come and say hi 👋 if you want to know how to proof if an LLM was trained on a data point!

[Min Choi @minchoi]

GPT-4o image gen is seriously impressive.

People are unlocking new creative ways to use it.

10 wild examples

[Parameter Lab @parameterlab]

👥 We're Hiring: Senior/Junior Data Engineer!

📍 Remote or Local | Full-Time or Part-Time

At https://t.co/oVYfRUtbdR, we’re building a platform that connects researchers and AI engineers worldwide—helping them stay ahead with daily digests, insightful summaries, and interactive events. Our LLM-powered ecosystem also bridges the gap between cutting-edge research and industry leaders. If you're passionate about data, AI, and making an impact, we’d love to have you on board!

What You’ll Do:
✔ Build Scalable Data Pipelines – Design and optimize workflows using tools like Airflow.
✔ Work Closely with AI Experts & Engineers – Collaborate to solve real-world data challenges.
✔ Optimize and Maintain Systems – Keep our data infrastructure fast, secure, and adaptable.

What You Bring:
✅ Proficiency in Airflow & PostgreSQL – You know your way around complex workflows and databases.
✅ Strong Python Skills – Clean, efficient, and maintainable code is your thing.
✅ (Bonus) Experience with LLMs – A huge plus as we integrate AI-driven solutions.
✅ Problem-Solving Mindset – You enjoy tackling challenges with real impact.
✅ Team Spirit – Excellent collaboration and communication.

Why Join Us?
🚀 Make a Difference – Your work directly enhances how research is shared and discovered.
🌍 Flexibility – Choose full-time or part-time, work remotely or locally.
⚡ Innovative Environment – AI, research, and data-driven solutions all in one place.
🤝 Great Team – Work with passionate, talented people shaping the future of research.

Ready to Join? Send your resume + a short note on why you’re a great fit to [email protected].
Be part of a team that’s redefining research with AI! #Hiring #DataEngineer #AI #RemoteJobs

[Parameter Lab @parameterlab]

🔎 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: https://t.co/JsXj0OTfnA

[Haritz Puerto @HaritzPuerto]

🧵 It is assumed that Membership Inference Attacks (MIA) do not work on LLMs, but our new paper shows it can work at the right scale! MIA is effective if the number of input tokens is large enough, such as in long documents and collections of them. 📃https://t.co/98JSkIw2OA

[Seong Joon Oh @coallaoh]

We just wanted to say: Membership inference is unlikely to succeed on n-grams or even paragraphs. Language models require **multiple documents** to gather enough evidence to determine whether specific data points were included in training. Accepted to #NAACL2025 Findings.

[Haritz Puerto @HaritzPuerto]

I'm excited to announce that my internship paper at @parameterlab was accepted to Findings of #NAACL2025 🎉 Huge thanks to @framart1 @coallaoh and @oodgnas! Amazing team!!

[Haritz Puerto @HaritzPuerto]

https://t.co/fVPEMUPn6M From time to time we hear news like this. However, proving that an LLM was trained on a specific document is very challenging 🥴 This motivated my latest work, where we show that current methods can be effective if we use enough data 🧐