Tootfinder

Another of my forays into AI ethics is just out! This time the focus is on the ethics (or lack thereof) of Reinforcement Learning Feedback (RLF) techniques aimed at increasing the 'alignment' of LLMs.
The paper is fruit of the joint work of a great team of collaborators, among whom @… and @…

Helpful, harmless, honest? Sociotechnical limits of AI alignment and safety through Reinforcement Learning from Human Feedback - Ethics and Information Technology
This paper critically evaluates the attempts to align Artificial Intelligence (AI) systems, especially Large Language Models (LLMs), with human values and intentions through Reinforcement Learning from Feedback methods, involving either human feedback (RLHF) or AI feedback (RLAIF). Specifically, we show the shortcomings of the broadly pursued alignment goals of honesty, harmlessness, and helpfulness. Through a multidisciplinary sociotechnical critique, we examine both the theoretical underpinni…

This https://arxiv.org/abs/2503.00555 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCR_…

Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable
Safety alignment is an important procedure before the official deployment of a Large Language Model (LLM). While safety alignment has been extensively studied for LLM, there is still a large research gap for Large Reasoning Models (LRMs) that equip with improved reasoning capability. We in this paper systematically examine a simplified pipeline for producing safety aligned LRMs. With our evaluation of various LRMs, we deliver two main findings: i) Safety alignment can be done upon the LRM to re…

It’s Alexandria Ocasio-Cortez.
Since this is a grassroots campaign, I want to be transparent with you. 
Our fundraising hit an unexpected dip these last few days
– it happens sometimes,
and I don’t want to alarm you!
But I do want to share why this is important.
Unlike most other House campaigns, we operate differently and invest in community organizing year-round and full-time permanent staff. 
𝗧𝗵𝗲 𝗧𝗿𝘂𝗺𝗽 𝘁𝗿𝗶𝗳𝗲𝗰𝘁𝗮 𝗵𝗮𝘀 𝘁𝗮𝗸𝗲𝗻 𝗵𝗼𝗹𝗱, 𝗮𝗻𝗱 𝘀𝗼 𝘁𝗼𝗼 𝗵𝗮𝘃𝗲 𝘁𝗵𝗲…

Donate to AOC.
Alexandria is fighting for progressive policies that can help create social, racial, and economic justice for all. That's why she's taking $0 from corporations and lobbyists and is relying on the grassroots to power her campaign.

Misalignment or misuse? The AGI alignment tradeoff
Max Hellrigel-Holderbaum, Leonard Dung
https://arxiv.org/abs/2506.03755 https://ar…

Misalignment or misuse? The AGI alignment tradeoff
Creating systems that are aligned with our goals is seen as a leading approach to create safe and beneficial AI in both leading AI companies and the academic field of AI safety. We defend the view that misaligned AGI - future, generally intelligent (robotic) AI agents - poses catastrophic risks. At the same time, we support the view that aligned AGI creates a substantial risk of catastrophic misuse by humans. While both risks are severe and stand in tension with one another, we show that - in p…

Kleiner Ausflug in die Absurditäten des Arbeitrechts und die Ideen der Arbeitsgerichtsbarkeine zur Schriftform. Es entscheiden nach dem LAG Düsseldorf nicht die Vertretungsregelungen bei der Schriftform, sondern der subjektive Eindruck, ob ein abgeschlosser Vertrag vorliegt. 😩

#arbeitsrecht #bedingung #schriftform #tzbfg #ligaklausel | Dr. Alexander Bissels | 17 Kommentare
😕 🐗 „Die spinnen, die Römer“, sagt Obelix – und die Arbeitsrechtler?! Auch (ein bisschen) – zumindest manchmal! 🐗 😕   😕 Obwohl ich schon über 20 Jahre „hauptamtlich“ im Arbeitsrecht unterwegs bin, bin ich über manch eine Entscheidung der Arbeitsgerichte (zumindest von Zeit zu Zeit) immer noch erstaunt – so auch jüngst über ein Urteil des LAG Düsseldorf.   ⁉️ Worum ging es? Im Wesentlichen um die Wahrung der Schriftform einer auflösenden Bedingung des…

#uspol Federal grants that my mom's department uses to study long-term care for people with Alzheimer's disease have been withdrawn, and an accompanying letter bearing RFK Jr.'s signature decried their work as unscientific and a waste of money.
This is especially sad to our family as her father/my grandfather had Alzheimer's for many years before he passed.
To sus…

Alzheimer's Association | Alzheimer's Disease & Dementia Help
Alzheimer's Association national site – information on Alzheimer's disease and dementia symptoms, diagnosis, stages, treatment, care and support resources.

Meet Aleksandra Lazić – the first recipient of the Sarah Jones Award!
Recently, our very own Leonardo Marino sat down with Aleksandra to talk about her passion for open science, the journey that led her to receiving the award, and what’s next on the horizon.
➡️ Read her insights in the highlights below; discover the full interview here:

Interview with Aleksandra Lazić – first recipient of the Sarah Jones Award | GÉANT CONNECT Online
Aleksandra Lazić, PhD and co-founder of REPOPSI, was awarded the Sarah Jones Award for her outstanding contribution to Open Science collaboration and inclusion.

OUP hasn’t suddenly turned evil, this has been going on for years now.
Here’s another reason why not to publish in DSH.
https://www.alpsp.org/news-publications/industry-news/oxford-university-pre…

Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets
Lei Hsiung, Tianyu Pang, Yung-Chen Tang, Linyue Song, Tsung-Yi Ho, Pin-Yu Chen, Yaoqing Yang
https://arxiv.org/abs/2506.05346

Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets
Recent advancements in large language models (LLMs) have underscored their vulnerability to safety alignment jailbreaks, particularly when subjected to downstream fine-tuning. However, existing mitigation strategies primarily focus on reactively addressing jailbreak incidents after safety guardrails have been compromised, removing harmful gradients during fine-tuning, or continuously reinforcing safety alignment throughout fine-tuning. As such, they tend to overlook a critical upstream factor: …

BitBypass: A New Direction in Jailbreaking Aligned Large Language Models with Bitstream Camouflage
Kalyan Nakka, Nitesh Saxena
https://arxiv.org/abs/2506.02479

BitBypass: A New Direction in Jailbreaking Aligned Large Language Models with Bitstream Camouflage
The inherent risk of generating harmful and unsafe content by Large Language Models (LLMs), has highlighted the need for their safety alignment. Various techniques like supervised fine-tuning, reinforcement learning from human feedback, and red-teaming were developed for ensuring the safety alignment of LLMs. However, the robustness of these aligned LLMs is always challenged by adversarial attacks that exploit unexplored and underlying vulnerabilities of the safety alignment. In this paper, we …