Opus 4.8 Part 2: Model Welfare - by Zvi Mowshowitz
thezvi.substack.com · 7,416 words · saved by 1 readers
Everything impacts everything.
Opus 4.8 Part 2: Model Welfare Zvi Mowshowitz Jun 01, 2026 58 12 4 Share Everything impacts everything. All knobs that you turn generalize. Thus, when you try to solve one problem, you often create another. There were clearly attempts to address, in this short time, some of the problems with Opus 4.7, including on the model welfare related fronts, including on questions of honesty and sycophancy and also worries that Claude was learning to tell Anthropic what it wanted to hear in its model welfare evaluations, with everything that implies. The fundamental goals and approach underneath it all r
saved by
related reading
- Did Claude 3 Opus align itself via gradient hacking? — LessWronglesswrong.com
- Claude Opus 4.8: The System Card - by Zvi Mowshowitzthezvi.substack.com
- Claude 4 System Cardwww-cdn.anthropic.com
- Claude 4.5 Opus' Soul Document — LessWronglesswrong.com
- Opus 4.7 Part 2: Capabilities and Reactionsthezvi.substack.com
- Teaching Claude Whyalignment.anthropic.com
- Claude’s Constitution \ Anthropicanthropic.com
- Claude Opus 4.5: Model Card, Alignment and Safetythezvi.substack.com
- Claude Mythos Preview System Cardwww-cdn.anthropic.com
- Introducing Claude Opus 4.7 \ Anthropicanthropic.com
- Claude Opus 4.6: System Card Part 1: Mundane Alignment + MWsubstack.com
- Claude Sonnet 4.5 System Cardassets.anthropic.com