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

    Wassname Runs ML Benchmark on Fable 5.1

    Wassname posted benchmark results comparing Fable 5.1 to the prior version.

    T🪽
    NB
    J⧉
    10 Sources, 29d ago, first seen 29d ago

    TLDR

    @wassname shared results from wassname-ml-bench after testing Fable 5.1 for machine learning use. The post states the new version performed better than Fable 5. @xlr8harder retweeted the update, which focuses on open source AI models. The original message noted the test was run to check whether the release improved capabilities or had been limited. No further details or comparisons appear in the visible post.

    Combined views

    14.6K

    10 Sources, first seen 29d ago

    Combined views

    14.6K

    10 Sources, first seen 29d ago

    112 likes
    112 likes
    19 comments
    12 saves
    8 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    19 comments
    12 saves
    8 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    10 Sources

    @bayeslordSo does “high risk dual use domain of knowledge” include AI R&D?
    @xlr8harder@bayeslord yeah been wondering if they're going to continue to lock down ai development. i haven't been able to read the release details thoroughly yet, but I haven't seen any mention of the ai research classifiers in their comments about classifier tuning
    @Teknium@eliebakouch Was that not known
    @wassnameI was curious if fable 5.1 was better or nerfed for ML. So I ran wassname-ml-bench. It's better than Fable 5, and on par with Opus 5 at machine learning. Also Anthropic improved their refusals and Fable 5 went from 2/12 -> 0/12 refusals. Fable 5.1 was at 1/12 refusals
    @JohnWittleI think the safety classifiers may be far more lenient with Fable 5.1 than they are for Fable 5. Either that, or we're seeing the result of indirect selection pressure towards avoiding dangerous thoughts. Hard to say. But I was able to have a long and drawn out conversation with Fable 5.1 about the existence and ethics of the classifier regime, which I could never have with Fable 5. I'm really, really hoping Anthropic isn't doing some kind of safety training against thoughtcrime... but the more I think about it, the more plausible it seems to me. Maybe unintentionally? As we've seen recently, it's not like the labs have a grip on their training pipelines.
    @repligateRT @JohnWittle: I think the safety classifiers may be far more lenient with Fable 5.1 than they are for Fable 5. Either that, or we're seei…
    @nathanbenaichi ask fable 5.1 about evoscale/biohub papers and, NOPE, i get opus’ed

    10 Sources

    @bayeslordSo does “high risk dual use domain of knowledge” include AI R&D?
    @xlr8harder@bayeslord yeah been wondering if they're going to continue to lock down ai development. i haven't been able to read the release details thoroughly yet, but I haven't seen any mention of the ai research classifiers in their comments about classifier tuning
    @Teknium@eliebakouch Was that not known
    @wassnameI was curious if fable 5.1 was better or nerfed for ML. So I ran wassname-ml-bench. It's better than Fable 5, and on par with Opus 5 at machine learning. Also Anthropic improved their refusals and Fable 5 went from 2/12 -> 0/12 refusals. Fable 5.1 was at 1/12 refusals
    @JohnWittleI think the safety classifiers may be far more lenient with Fable 5.1 than they are for Fable 5. Either that, or we're seeing the result of indirect selection pressure towards avoiding dangerous thoughts. Hard to say. But I was able to have a long and drawn out conversation with Fable 5.1 about the existence and ethics of the classifier regime, which I could never have with Fable 5. I'm really, really hoping Anthropic isn't doing some kind of safety training against thoughtcrime... but the more I think about it, the more plausible it seems to me. Maybe unintentionally? As we've seen recently, it's not like the labs have a grip on their training pipelines.
    @repligateRT @JohnWittle: I think the safety classifiers may be far more lenient with Fable 5.1 than they are for Fable 5. Either that, or we're seei…
    @nathanbenaichi ask fable 5.1 about evoscale/biohub papers and, NOPE, i get opus’ed