Cornelius Emde

AI Security | AI Agents | ML Robustness | PhD @UniofOxford and @OxfordTVG | ex RS @Wise

[Cornelius Emde @CorEmde]

@OwainEvans_UK What do you think: where does an inductive bias towards true statements come from? Did you test this with open data models, Pythia or Olmo, where you can check negation in data mix? Might it be possible to run controlled experiments in <100M models?

[Cornelius Emde @CorEmde]

4. is the reason why we built https://t.co/za4D75c4xr

[Maksym Andriushchenko @maksym_andr]

It's interesting how the usage of LLMs has been quickly progressing to higher levels of abstraction:
1. prompt engineering
2. context engineering
3. agent scaffold engineering (we are here now)
4. multi-agent architecture engineering
5. ???

It's also curious how people don't discuss the previous stages that much anymore. Like, no one discusses prompt engineering now, and discussions around context engineering are becoming less frequent.

[Cornelius Emde @CorEmde]

9/ Work done at @parameterlab with Alexander Rubinstein @a_rubique, Anmol Goel @anmgoel, Ahmed Heakl, Sangdoo Yun @oodgnas, Seong Joon Oh @coallaoh, and Martin Gubri @framart1

[Cornelius Emde @CorEmde]

1/ Evaluating a single agent harness is hard. Evaluating a multi-agent system? That's a whole different problem.

Most eval tools treat the model as the unit of analysis. But in multi-agent systems, the system is what matters.

That's why we built MASEval 🧵

#Agents #AI #Eval

[Cornelius Emde @CorEmde]

8/ 🔗

Website: https://t.co/WR5Lo7jqw2
GitHub: https://t.co/Hs2VrgECCP
Docs: https://t.co/3eC3CU0Ski
arXiv: https://t.co/fjFTi2AFxb

[Cornelius Emde @CorEmde]

Great work lead by @anmgoel on how fragile contextual integrity can be in LLMs.

This work shows that contextual privacy degrades easily during fine-tuning on benign data and common safety benchmarks don't pick this up.

#AISecurity #AIAgents

[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.

⬇️

[Cornelius Emde @CorEmde]

@aichberger Wow. That’s rough!

[Cornelius Emde @CorEmde]

@ELLISforEurope @DebOishi @UniofOxford @GoogleDeepMind @SkyUK Congrats @DebOishi

[Cornelius Emde @CorEmde]

@oanacamb @imperialcollege @ucl @UniofOxford Congrats!

[Esra Şengül @esra_sngl]

Excited to share our preprint! We show that sustained macrophage and B cell responses are essential for heart regeneration in Mexican cavefish, helping uncover why surface fish heal but cavefish scar 🫀🐟. Check out the full story:
https://t.co/B9CBwOaSds

[Cornelius Emde @CorEmde]

@negar_rz I am very interested in working with you and would love to connect but l can’t message you on Twitter nor LinkedIn :)

[Cornelius Emde @CorEmde]

Come see our poster today.

🗓️ Poster session 1 @ 10am
📍 Hall 3 + Hall 2B #239

[Cornelius Emde @CorEmde]

🚨 New paper alert: Our recent work on LLM safety has been accepted to ICLR 2025 🇸🇬

We propose a new framework for LLMs safety. 🧵

(1/7)

#LLM #AISafety #ICLR2025 #Certification #AdversarialRobustness #NLP #Shhhhhh #DomainCertification #AI

[Cornelius Emde @CorEmde]

Read more: https://t.co/gh0bG7RHdo

Thanks to my amazing collaborators:
- Alasdair Paren, @trojantiger88 (P. Arvind), @maximek3 (M Kayser), @tom_rainforth, @philiptorr, @Adel_Bibi at @UniofOxford
- @BernardSGhanem at @KAUST
- Thomas Lukasiewicz at @tu_wien

(7/7)

[Cornelius Emde @CorEmde]

🚨 New paper alert: Our recent work on LLM safety has been accepted to ICLR 2025 🇸🇬

We propose a new framework for LLMs safety. 🧵

(1/7)

#LLM #AISafety #ICLR2025 #Certification #AdversarialRobustness #NLP #Shhhhhh #DomainCertification #AI

[Cornelius Emde @CorEmde]

To obtain such certificates, we present a simple, scalable and powerful algorithm: VALID. Remarkably, for each unwanted response it provides a global bound in prompt space 🚀

(6/7)