Gary Marcus

Rank 12 of 47
|
Score 83
5/18/2024, 12:09:18 AM
X
In reply to:
Leonardo
@developer134135
·
613d
@GaryMarcus Because it doesn't have space in the parameter set. There are a tiny number of chess moves in the pre-training dataset and none in the fine-tuning set.

Someone already did a project where they fine-tuned an LLM on chess games, and yes, it knows the game and plays excellently.
Gary Marcus
@GaryMarcus
·
613d
Absolutely stunning: anybody who has ever played chess knows you can’t take your own pieces.

But ChatGPT doesn’t. Even after being trained on millions of games.

Incredible.
Abhijit Mahabal
@amahabal
·
613d
@GaryMarcus Similarly with chess moves. “Chess move” it learns, but even the category “legal chess move” stumps it, as shown by this amazing video https://www.reddit.com/r/Anarchy… of ChatGPT playing Stockfish, where it appears unconstrained by legality, happy even to take its own material. 5/n

The statement is part of a technical discussion about the capabilities and limitations of a language model (ChatGPT) in understanding and executing chess moves legally. The conversation involves multiple participants discussing specific aspects of AI training and its application to chess, highlighting issues with the model's understanding of legal versus illegal moves. This discussion is technical and pertains to the field of artificial intelligence and machine learning, specifically focusing on the application of these technologies in game scenarios like chess.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does no harm as it is a factual critique of a technical work. [+1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The statement respects privacy and dignity, focusing solely on the technical aspects of the work without any personal attacks. [+1]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement promotes understanding by sharing insights about the limitations of current AI technology in understanding chess. [+1]
  4. 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 criticism by pointing out flaws in a previous project without demeaning language or personal attacks. [+1]
  5. Principle 6:
    I will use my influence for the betterment of society.
    The statement uses influence to better society by contributing to a deeper understanding of AI capabilities and limitations, which is crucial for technological advancement. [+1]
  6. Principle 7:
    I will uphold the principles of free speech and use my platform responsibly and with integrity.
    The statement upholds the principles of free speech responsibly, contributing positively to a technical discussion without misinformation or harmful content. [+1]