Gary Marcus

Rank 15 of 47

Score 76

@thegartsy This is actually an LLM combined with a classical symbolic search mechanism. an LLM on its own could not do this.

6/16/2024, 4:26:20 AM

In reply to:

Tess Hegarty

@thegartsy

512d

I’ve seen people compare LLMs to calculators, or search, or call it just mimicry. That seems so disingenuous to me, relative to all the different benchmarks LLMs have achieved.

Deedy

@deedydas

513d

It's finally here. Q* rings true. Tiny LLMs are as good at math as a frontier model.

By using the same techniques Google used to solve Go (MTCS and backprop), Llama8B gets 96.7% on math benchmark GSM8K!

That’s better than GPT-4, Claude and Gemini, with 200x less parameters!

The statement by @thegartsy engages in a technical discussion about the capabilities and mechanisms of language models (LLMs), specifically addressing the comparison of LLMs to calculators or search mechanisms. The reply by @thegartsy clarifies the nature of the LLM's functionality, emphasizing the combination of LLMs with classical symbolic search mechanisms. This contributes to the public discourse on the understanding and evaluation of AI technologies.

Principle 1:
I will strive to do no harm with my words and actions.
The statement strives to do no harm and is focused on providing a factual clarification about the technology, which aligns with the principle of doing no harm. [+1]
Principle 3:
I will use my words and actions to promote understanding, empathy, and compassion.
The statement promotes understanding by explaining the technical aspects of LLMs, thereby fostering a more informed discussion. [+2]
Principle 4:
I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
The statement engages in constructive dialogue by addressing a point made in the conversation without resorting to personal attacks or ad hominem arguments. [+1]