Will a major AI lab claim to use activation steering in its main chat assistant by EOY 2025?
Basic
13
Ṁ5172026
31%
chance
1D
1W
1M
ALL
Also includes methods inspired by activation steering, as long as they don't use any gradient descent step.
Only includes announcements about main chat assistants (e.g. Claude, ChatGPT, Bard, ...) of a major AI lab (OpenAI, Google Deepmind, Anthropic, Meta, Inflection or Mistral).
Does not include to fine-tuning API endpoints.
This question is managed and resolved by Manifold.
Get
1,000
and3.00
Sort by:
Anthropic is running a demo of an activation-steered Claude obsessed with the Golden Gate Bridge: https://www.reddit.com/r/singularity/comments/1cz7kuh/claude_golden_gate_bridge_is_now_available_bridge/ (Context: https://www.anthropic.com/research/mapping-mind-language-model )
Related questions
Related questions
Will OpenAI hint at or claim to have AGI by 2025 end?
29% chance
Will xAI AI be a Major AI Lab by 2025?
24% chance
Will there be another blatant demonstration of AI risks, comparable to Bing Chat, by 2024?
30% chance
Will Microsoft's new AI research team release any new product by EOY 2024?
32% chance
Will a consumer-grade autonomous AI be released by an established tech firm by the end of 2024?
31% chance
Will OpenAI announce a major breakthrough in AI alignment in 2024?
21% chance
Who will have the most popular AI assistant at the end of 2025? (judged by active users)
Will Anthropic announce one of their AI systems is ASL-3 before the end of 2025?
59% chance
Will Meta's Threads have an AI chatbot by 2025?
67% chance
Will a major AI company acknowledge the possibility of conscious AIs by 2026?
85% chance