Is Nick Cammarata right: LLM will be able to mechanistically audit own circuits and explain ghiblification in 2Y? | Manifold

Is Nick Cammarata right: LLM will be able to mechanistically audit own circuits and explain ghiblification in 2Y?

6

100Ṁ109

2027

34%

chance

1D

1W

1M

ALL

Background

This is based on this Nick Cammarata tweet thread.

https://x.com/nickcammarata/status/1905321653401518517?s=46

“I’m at like 80% that in two years we’ll be able to get 4o to mechanistically audit its own circuits and explain what the core ideas behind ghiblification are“

Resolution Criteria

I will try to ask Nick on Twitter or find another sensible resolution.

Mechanistic interpretability

Get

1,000

to start trading!

Sort by:

Can you specify a bit more in detail what level of detail of the explanation must be to resolve yes? Is it enough to give generic description? ("A neutral network of this and this architecture needs to be train on this data")? Does the LLM need to be able to write a program with no training data that can do the job? (The other extreme)

Related questions

Are LLMs capable of reaching AGI?

LLM Hallucination: Will an LLM score >90% on SimpleQA before 2026?

Will LLM based systems have debugging ability comparable to a human by 2030?

LLMs by EOY 2025: Will Retentive Learning Surpass Transformers? (Subsidised 400 M$)

Will there be major breakthrough in LLM Continual Learning before 2026?

Will LLMs be able to formally verify non-trivial programs by the end of 2025?

When will the next paradigm in LLMs (after reasoning) be released?

Will one of the major LLMs be capable of continual lifelong learning (learning from inference runs) by EOY 2025?

Will LLM inference for the largest models run on analogue circuitry as the primary element of computuation by end 2028?

EOY 2025: Will open LLMs match closed-source LLMs on coding to within 50 ELO points?

Related questions

Are LLMs capable of reaching AGI?

Will LLMs be able to formally verify non-trivial programs by the end of 2025?

LLM Hallucination: Will an LLM score >90% on SimpleQA before 2026?

When will the next paradigm in LLMs (after reasoning) be released?

Will LLM based systems have debugging ability comparable to a human by 2030?

Will one of the major LLMs be capable of continual lifelong learning (learning from inference runs) by EOY 2025?

LLMs by EOY 2025: Will Retentive Learning Surpass Transformers? (Subsidised 400 M$)

Will LLM inference for the largest models run on analogue circuitry as the primary element of computuation by end 2028?

Will there be major breakthrough in LLM Continual Learning before 2026?

EOY 2025: Will open LLMs match closed-source LLMs on coding to within 50 ELO points?

© Manifold Markets, Inc.•Terms + Mana-only Terms•Privacy•Rules