Will an uncensored open-source LLM model comparable to 2023 GPT4 be available to the public by the end of 2025?

410Ṁ469

Feb 18

84%

chance

ALL

Question resolves to YES if, by the end of 2025, there is a publically available and open-source LLM that can be used to generate uncensored text and that has performance comparable to GPT4 as it performed in early 2023.

If there are no public open-source LLMs comparable to 2023 GPT4 or the ones available can only generate content that is censored to some degree, the question resolves to NO.

Technology

Technical AI Timelines

LLMs

Get

1,000

to start trading!

People are also trading

Will we have an open-source model that is equivalent GPT-4 by end of 2025?

82% chance

Which next-gen frontier LLMs will be released before GPT-5? (2025)

Will an open-source fully functional Auto-GPT like LLM exist by the end of 2025?

90% chance

Will an open-source LLM beat or match GPT-4 by the end of 2024?

83% chance

Will a Mamba-based LLM of GPT 3.5 quality or greater be open sourced in 2024?

73% chance

Will a model as great as GPT-5 be available to the public in 2025?

84% chance

Will any open source LLM with <20 billion parameters outperform GPT-4 on most language benchmarks by the end of 2024?

13% chance

When will an open-source LLM be released with a better performance than GPT-4?

Will there be an LLM (as good as GPT-4) that was trained with 1/100th the energy consumed to train GPT-4, by 2026?

82% chance

an LLM as capable as GPT-4 runs on one 4090 by March 2025

Sort by:

Uh, you could specialize a model on a specific topic and beat gpt4 but what about other fields?
AFAIK beating gpt's funding on training is pretty hard and partly why they got the lead.

Roughly speaking Yes but this would be impossible to judge accurately unless OpenAI provides long term access to the OG GPT-4 model, which is unlikely.

How do you measure "performance"?

Depending on how you specify that, it may be possible to resource this market YES today.

@JosefMitchell I would use at least one open-source benchmarking tool, such as GPT-Fathom (https://arxiv.org/abs/2309.16583) to compare the benchmark scores to another LLM that was evaluated in a similar manner. I would also accept credible papers that vouched for the capabilities of an LLM and that compared them favorably to GPT-4. I may accept public statements from AI researchers about the performance and capabilities of an open-source LLM provided these statements would be testable and I could evaluate their claims.