Gary Marcus

Rank 11 of 47
|
Score 114
10/4/2024, 11:13:30 PM
X
In reply to:
Benjamin Riley
@benjaminjriley
·
314d
@GaryMarcus An interesting question here is how far the mitigation of LLM limitations can go. Both this paper and my own usage indicate it is getting harder to "outwit" these tools.
Gary Marcus
@GaryMarcus
·
314d
‼️Devastating new result showing the same old thing: “o1 still displays the same qualitative trends that we observed in previous systems. Specifically, o1 - like previous LLMs - is sensitive to the probability of examples and tasks, performing better and requiring fewer "thinking…

The statement is part of a technical discussion about the limitations of large language models (LLMs), specifically their sensitivity to low-frequency outliers. It contributes to public discourse by engaging with ongoing debates about the capabilities and limitations of AI models.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does no harm and is a neutral observation about AI limitations. [+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    It promotes understanding by discussing specific technical challenges in AI. [+1]
  3. 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 constructively with the ongoing discussion about AI limitations without personal attacks. [+1]