• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Paper Links Tokenization to Cross-Lingual Transfer Failures

    alphaXiv post summarizes paper on pretraining failures in cross-lingual LLM knowledge transfer.

    AL
    1 Source, 32d ago, first seen 32d ago

    TLDR

    The official alphaXiv account posted about the paper “Why Pretraining Fails to Share Cross-Lingual Knowledge.” The post states that knowledge learned in one language sometimes barely transfers to another. It reports that the failure begins during pretraining and that disjoint token spaces alone can cause the issue, even when comparing two otherwise identical copies of English. The account also generated the headline “Study Shows Tokenization Bottleneck Limits Cross-Ling…” No further details or independent confirmations appear in the packet.

    Combined views

    19.1K

    1 Source, first seen 32d ago

    Combined views

    19.1K

    1 Source, first seen 32d ago

    302 likes
    302 likes
    12 comments
    217 saves
    43 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    12 comments
    217 saves
    43 reposts
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    1 Source

    @askalphaxiv“Why Pretraining Fails to Share Cross-Lingual Knowledge” LLM's knowledge learned in one language Sometimes barely transfers to another. This paper identifies that the failure starts during pretraining and that disjoint token spaces alone can cause it, even between two otherwise identical copies of English. Mapping languages into a shared semantic token space improves cross-lingual learning by up to 14x, suggesting tokenization itself is a major bottleneck to truly shared multilingual knowledge. https://www.alphaxiv.org/abs/2608.pretraining-fails-cross-lingual-knowledge

    1 Source

    @askalphaxiv“Why Pretraining Fails to Share Cross-Lingual Knowledge” LLM's knowledge learned in one language Sometimes barely transfers to another. This paper identifies that the failure starts during pretraining and that disjoint token spaces alone can cause it, even between two otherwise identical copies of English. Mapping languages into a shared semantic token space improves cross-lingual learning by up to 14x, suggesting tokenization itself is a major bottleneck to truly shared multilingual knowledge. https://www.alphaxiv.org/abs/2608.pretraining-fails-cross-lingual-knowledge