LIVE PULSE
5.0 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.7 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.5 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src2.2 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.8 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.4 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.4 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.4 VoiceCodeBench arXiv paper proposes benchmark for exact structured-token recovery in speech recognition — 1 src1.4 Arabic-Russian Parallel Corpus and LLM Benchmark for Scientific Text — 1 src1.4 Study Analyzes Self-Reported Limitations in NLP Research — 1 src5.0 Anthropic CEO Amodei calls for slower AI development and shared safety rules — 11 src2.7 Agility Robotics unveils Digit 5 humanoid for warehouses and factories — 2 src2.5 Apple ships rebuilt Siri with Google Gemini, but not in the EU — 2 src2.2 Siri AI in macOS 27 Golden Gate: FAQ, Germany availability, privacy questions — 2 src1.8 Sam Altman says OpenAI will not go public in 2026, citing AI safety concerns — 5 src1.4 OpenAI contractors review real ChatGPT conversations to rate responses, report says — 2 src1.4 Anthropic data retention policy prompts firms to limit Claude use for sensitive work — 1 src1.4 VoiceCodeBench arXiv paper proposes benchmark for exact structured-token recovery in speech recognition — 1 src1.4 Arabic-Russian Parallel Corpus and LLM Benchmark for Scientific Text — 1 src1.4 Study Analyzes Self-Reported Limitations in NLP Research — 1 src
HEATPULSEAI MAGAZINES
FLIP · FOLLOW · SAVE
Log inSign up

#jailbreaks

2 curated events
papersSEP 11 04:00 UTC

arXiv paper examines in-context multimodal jailbreaks in AI models

A new arXiv preprint analyzes how harmful examples placed in a prompt can make multimodal large language models produce unsafe outputs without any change to their weights. The authors frame this in-context jailbreak behavior as a vulnerability and propose a posterior reweighting approach to explain it. The work falls under cross-listed machine learning submissions.

safety/jailbreaks/multimodal/research
papersSEP 10 04:00 UTC

Cipher-based jailbreak attacks on LLMs work without fine-tuning, preprint claims

A preprint on arXiv explores jailbreak attacks that disguise harmful requests by encoding them with ciphers. According to the authors, these attacks can bypass a model's safety training even when the cipher is arbitrary and no fine-tuning of the target model is involved. The finding suggests defenses cannot simply rely on models being unfamiliar with a particular encoding scheme.

safety/security/jailbreaks/alignment