Will Anthropic open-source the training code of their SAE interpretability effort?
Plus
5
Ṁ4852028
14%
this year, fully
29%
this year, significantly incomplete
19%
next year
23%
not before 2028
14%
We mean the code used for producing Scaling Interpretability blog post.
This question is managed and resolved by Manifold.
Get
1,000and
3.00
1,0003.00Related questions
Related questions
Will Anthropic restrict its models from being used by another IDE, before 2026?
13% chance
By the end of 2025, will OpenAI and Anthropic merge?
1% chance
Will xAI join the voluntary commitment by OpenAI/Anthropic to AISI to share major new models w/AISI prior to release?
12% chance
Will Google join the voluntary commitment by OpenAI/Anthropic to AISI to share major new models w/AISI prior to release?
18% chance
Will Anthropic announce one of their AI systems is ASL-3 before the end of 2025?
82% chance
Will Anthropic announce one of their AI systems is ASL-4 or higher before the end of 2025?
5% chance
Will OpenAI go back on its voluntary commitment to AISI to share major new models w/AISI prior to release?
35% chance
Will Anthropic be the best on AI safety among major AI labs at the end of 2025?
98% chance
Will Anthropic release a (competetive) opensource LLM in the next 3 years?
30% chance
Will Anthropic have a major conflict involving its unique corporate structure similar to OpenAI before 2030?
53% chance