Will someone create a LLM/chatbot fine-tuned to Eliezer's LessWrong sequences by EOY2025?
Basic
2
Ṁ402026
66%
chance
1D
1W
1M
ALL
This question is managed and resolved by Manifold.
Get
1,000and
3.00
1,0003.00Sort by:
Apparently there's LessWrong-Amplify-Instruct, an amplified dataset for instruction tuning based on LW posts. It would seem a lot of other models have been fine-tuned using that dataset, particularly Nous's Capybara family of models.
That said, I haven't found any LLMs purpose-finetuned just for LW/rat content
Related questions
Related questions
Who will have the best LLM at the end of 2025 (as decided by ChatBot Arena)?
Who will have the best LLM at the end of February (as decided by ChatBot Arena)?
Meta to release world's most advanced AI?
15% chance
Will the most interesting AI in 2027 be a LLM?
52% chance
Will Quora questions be auto-answered by a more sophisticated bot (at level of GPT3.5 or higher) by EOY 2025?
77% chance
Will chatbots/AI be powerful enough to make me unsad by EOY2025?
37% chance
Will there be an LLM which can do fluent conlang translations by EOY 2024?
72% chance
By the end of 2024, will any LessWrong, EA, or Alignment forum post with at least 100 upvotes be revealed to have been written by an AI?
21% chance
Will any LLM be able to consistently play Akinator correctly as the user by 2028?
79% chance
Will a chatbot from a Chinese company top the LMSYS leaderboard in 2026?
20% chance