Lightnews — Scholar-powered news

Reposted by Maksym Andriushchenko

Johannes Zenn @johanneszenn.bsky.social · 27d

A new recording of our FridayTalks@Tübingen series is online!

AI Safety and Alignment
by
@maksym-andr.bsky.social

Watch here: youtu.be/7WRW8MDQ8bk

AI Safety and Alignment - [Maksym Andriushchenko]

YouTube video by Friday Talks Tübingen

youtu.be

1 1 6

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

Thank you so much, Nicolas! :)

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

We believe getting this—some may call it "AGI"—right is one of the most important challenges of our time.

Join us on this journey!

11/11

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

Taking this into account, we are only interested in studying methods that are general and scale with intelligence and compute. Everything that helps to advance their safety and alignment with societal values is relevant to us.

10/n

1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

Broader vision. Current machine learning methods are fundamentally different from what they used to be pre-2022. The Bitter Lesson summarized and predicted this shift very well back in 2019: "general methods that leverage computation are ultimately the most effective".

9/n

1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

... —literally anything that can be genuinely useful for other researchers and the general public.

8/n

1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

Research style. We are not necessarily interested in getting X papers accepted at NeurIPS/ICML/ICLR. We are interested in making an impact: this can be papers (and NeurIPS/ICML/ICLR are great venues), but also open-source repositories, benchmarks, blog posts, even social media posts ...

7/n

1 1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

For more information about research topics relevant to our group, please check the following documents:
- International AI Safety Report,
- An Approach to Technical AGI Safety and Security by DeepMind,
- Open Philanthropy’s 2025 RFP for Technical AI Safety Research.

6/n

1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

We're also interested in rigorous AI evaluations and informing the public about the risks and capabilities of frontier AI models. Additionally, we aim to advance our understanding of how AI models generalize, which is crucial for ensuring their steerability and reducing associated risks.

5/n

1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

Research group. We will focus on developing algorithmic solutions to reduce harms from advanced general-purpose AI models. We're particularly interested in alignment of autonomous LLM agents, which are becoming increasingly capable and pose a variety of emerging risks.

4/n

1 2

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

Please fill the Google form forms.gle/GJB9bp7gRXAf... if you're interested!

3/n

Expression of Interest: AI Safety and Alignment Group (Maksym Andriushchenko)

Hi 👋! My name is Maksym Andriushchenko. I'm starting my research group at the ELLIS Institute Tübingen and Max Planck Institute for Intelligent Systems in September 2025 that focuses on AI safety and ...

forms.gle

1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

Hiring. I'm looking for multiple PhD students: both those able to start in Fall 2025 and through centralized programs like CLS, IMPRS, and ELLIS (the deadlines are in November) to start in Spring–Fall 2026. I'm also searching for postdocs, master's thesis students, and research interns.

2/n

1

Maksym Andriushchenko @maksym-andr.bsky.social · Aug 6

🚨 Incredibly excited to share that I'm starting my research group focusing on AI safety and alignment at the ELLIS Institute Tübingen and Max Planck Institute for Intelligent Systems in September 2025! 🚨

1/n

2 9

Maksym Andriushchenko @maksym-andr.bsky.social · Jun 19

This is joint work with amazing collaborators: Thomas Kuntz, Agatha Duzan, Hao Zhao, Francesco Croce, Zico Kolter, and Nicolas Flammarion.

It will be presented as an oral at the WCUA workshop at ICML 2025!

Paper: arxiv.org/abs/2506.14866
Code: github.com/tml-epfl/os-...

GitHub - tml-epfl/os-harm: OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents

OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents - tml-epfl/os-harm

github.com

2

Maksym Andriushchenko @maksym-andr.bsky.social · Jun 19

Main findings based on frontier LLMs:
- They directly comply with _many_ deliberate misuse queries
- They are relatively vulnerable even to _static_ prompt injections
- They occasionally perform unsafe actions

1

Maksym Andriushchenko @maksym-andr.bsky.social · Jun 19

🚨Excited to release OS-Harm! 🚨

The safety of computer use agents has been largely overlooked.

We created a new safety benchmark based on OSWorld for measuring 3 broad categories of harm:
1. deliberate user misuse,
2. prompt injections,
3. model misbehavior.

1 2 3

Maksym Andriushchenko @maksym-andr.bsky.social · Dec 9

Paper link: josh-freeman.github.io/resources/ny....

This is joint work with amazing collaborators: Joshua Freeman, Chloe Rippe, and Edoardo Debenedetti.

🧵3/n

josh-freeman.github.io

6

Maksym Andriushchenko @maksym-andr.bsky.social · Dec 9

3. However, the memorized articles cited in the NYT lawsuit were clearly cherry-picked—random NYT articles have not been memorized.

4. We also provide a legal analysis of this case in light of our findings.

We will present this work at the Safe Gen AI Workshop at NeurIPS 2024 on Sunday.

🧵2/n

1 6

Maksym Andriushchenko @maksym-andr.bsky.social · Dec 9

🚨Excited to share our new work!

1. Not only GPT-4 but also other frontier LLMs have memorized the same set of NYT articles from the lawsuit.

2. Very large models, particularly with >100B parameters, have memorized significantly more.

🧵1/n

1 1 15

Maksym Andriushchenko @maksym-andr.bsky.social · Dec 7

3. Exploring Memorization and Copyright Violation in Frontier LLMs: A Study of the New York Times v. OpenAI 2023 Lawsuit (openreview.net/forum?id=C66...)

Thanks to the amazing collaborators for making all these works possible!

🧵4/4

Exploring Memorization and Copyright Violation in Frontier LLMs: A...

Copyright infringement in frontier large language models (LLMs) has received much attention recently due to the case NYT v. Microsoft, filed in December 2023. The New York Times claims that GPT-4...

openreview.net

Maksym Andriushchenko @maksym-andr.bsky.social · Dec 7

3. Improving Alignment and Robustness with Circuit Breakers (arxiv.org/abs/2406.04313)

Workshops:
1. Does Refusal Training in LLMs Generalize to the Past Tense? (arxiv.org/abs/2407.11969)
2. Is In-Context Learning Sufficient for Instruction Following in LLMs? (arxiv.org/abs/2405.19874)

🧵3/4

Improving Alignment and Robustness with Circuit Breakers

AI systems can take harmful actions and are highly vulnerable to adversarial attacks. We present an approach, inspired by recent advances in representation engineering, that interrupts the models as t...

arxiv.org

1 1

Maksym Andriushchenko @maksym-andr.bsky.social · Dec 7

Here are the 6 papers that we are presenting at NeurIPS 2024:

Main track:
1. Why Do We Need Weight Decay in Modern Deep Learning? (arxiv.org/abs/2310.04415)
2. JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models (arxiv.org/abs/2404.01318)

🧵2/4

Why Do We Need Weight Decay in Modern Deep Learning?

Weight decay is a broadly used technique for training state-of-the-art deep networks from image classification to large language models. Despite its widespread usage and being extensively studied in t...

arxiv.org

1

Maksym Andriushchenko @maksym-andr.bsky.social · Dec 7

📢 I'll be at NeurIPS 🇨🇦 from Tuesday to Sunday!

Let me know if you're also coming and want to meet. Would love to discuss anything related to AI safety/generalization.

Also, I'm on the academic job market, so would be happy to discuss that as well! My application package: andriushchenko.me.

🧵1/4

Maksym Andriushchenko

I'm a PhD student in computer science at EPFL advised by Nicolas Flammarion. I'm interested in understanding why machine learning works and why it fails.

andriushchenko.me

1 5

Reposted by Maksym Andriushchenko

Marcel Salathé @marcelsalathe.bsky.social · Dec 6

Mindblowing: EPFL PhD student @maksym-andr.bsky.social, winner of best CS thesis award, showed that leading hashtag#AI models are not robust to simple adaptive jailbreaking attacks. Indeed, he managed to jailbraik all models with a 100% success rate 🤯

Jailbraking paper: arxiv.org/abs/2404.02151

Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks

We show that even the most recent safety-aligned LLMs are not robust to simple adaptive jailbreaking attacks. First, we demonstrate how to successfully leverage access to logprobs for jailbreaking: we...

lnkd.in

1 12

Maksym Andriushchenko @maksym-andr.bsky.social · Nov 20

really feels like Twitter circa 2018. good old days... 😀

10