5Don't Worry About the Vase (Zvi Mowshowitz)·
Zvi Mowshowitz commentary on Anthropic's 'Verbalizable Representations Form a Global Workspace in Language Models'
Zvi Mowshowitz highlights a new Anthropic paper titled 'Verbalizable Representations Form a Global Workspace in Language Models,' describing it as 'very cool.' The post links to both the paper and an Anthropic blog post version. The underlying paper appears to investigate how language models form internal representations that can be verbalized, connecting to global workspace theory from cognitive science.
Related guides (4)
Related events (8)
Opus 4.7 Part 3: Model Welfare
Zvi Mowshowitz publishes a commentary piece on model welfare in the context of Anthropic's Claude Opus 4.7, crediting Anthropic for enabling the discussion. The piece appears to engage with questions about the moral status or wellbeing of AI models. As a tier-2 commentary source, this reflects ongoing discourse in the AI safety and alignment community about how to think about model welfare as frontier models grow more capable.
What is Anthropic?
A commentary piece from Zvi Mowshowitz's 'Don't Worry About the Vase' analyzing Anthropic as a company. The piece appears to examine Anthropic's identity, mission, and strategic positioning. As a Tier 2 source commentary on a major AI safety lab, it likely covers Anthropic's stated goals around safety-focused AI development and its commercial trajectory.
Opus 4.8 Part 2: Model Welfare
Zvi Mowshowitz publishes a commentary piece on model welfare in the context of Claude Opus 4.8, continuing a multi-part analysis. The piece appears to engage with questions about AI moral status and welfare considerations as they relate to Anthropic's latest model. The body content is minimal in the provided excerpt, but the topic sits squarely within ongoing AI safety and alignment discourse.
Why Language Models Hallucinate
OpenAI published research explaining the mechanisms behind language model hallucination. The work connects improved evaluation methods to enhanced AI reliability, honesty, and safety. The body is sparse on technical detail, but the framing positions this as foundational research relevant to alignment and deployment trust.
Anthropic demonstrates feature steering in Claude 3 Sonnet via interpretability research
Anthropic released a 24-hour public demo called 'Golden Gate Claude' to illustrate findings from a major interpretability paper on Claude 3 Sonnet. The research identifies millions of internal 'features' — neuron combinations that activate for specific concepts — and shows these can be surgically amplified or suppressed to alter model behavior without prompting or fine-tuning. The Golden Gate Bridge feature was amplified as a demonstration, causing the model to reference the bridge in nearly all responses. Anthropic argues this mechanistic control over internal activations has direct implications for AI safety, including the ability to modulate safety-relevant features like those tied to deception or dangerous code.
Survey chapter on LLM mechanisms, emergent capabilities, and cognition debates
A new arXiv preprint surveys current understanding of large language models, covering the Transformer architecture, emergent capabilities resembling human cognition (symbolic reasoning, theory of mind, deception), and explainability approaches from neuron activation analysis to circuit tracing. The chapter also engages the debate over whether LLMs genuinely understand or merely pattern-match, arguing against reductive anti-anthropomorphism while acknowledging human-LLM differences. It is framed as a book chapter synthesizing recent empirical findings and theoretical positions.
Vision Language Models Explained
A Hugging Face blog post providing a technical overview of vision language models (VLMs), covering their architecture, training approaches, and capabilities. The post serves as an educational resource explaining how VLMs combine visual and language understanding. As a tier-2 commentary piece, it synthesizes existing knowledge rather than presenting new research findings.
VLMs overestimate common ground in asymmetric dialogue, conflating potential with established shared understanding
A new arXiv paper investigates whether vision-language models can distinguish between what could be shared versus what has actually been established as shared between dialogue participants. Using 13,077 annotated reference expressions from HCRC MapTask dialogues, the authors find that VLMs systematically over-predict alignment when given task-relevant map content—whether presented visually or as text—suggesting the bias stems from static referential cues rather than tracking grounding through dialogue history. The effect is observed most strongly in Qwen3-VL-8B-Instruct and replicated across four additional models from two architecture families, revealing a fundamental limitation in how current VLMs model collaborative dialogue.