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

Looped small decision model reportedly scores 72.0% versus 58.4% for a nonlooped counterpart

A post says SanSi runs Ouro-1.4B for up to eight loops and can answer after any loop.

1 Source, 1h ago, first seen 1h ago

TLDR

A post summarizing a paper says SanSi repeatedly runs the layers of Ouro-1.4B and checks option probabilities after each of up to eight loops. Across 10,027 test decisions, it reports 72.0% accuracy versus 58.4% for a same-shape model without loops, at 7.7 times the compute. On liar chains longer than any seen in training, the post says SanSi reached 72.6%, compared with 50.0% for a 4.2B single-pass model.

Combined views

—

1 Source, first seen 1h ago

— likes— comments— saves— reposts

Combined views

—

1 Source, first seen 1h ago

— likes— comments— saves— reposts

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Featured Source

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

1 Source

Rohan Paul@rohanpaul_aiThis paper finds that small decision models get far more accurate by rerunning the same layers several times before giving an answer. Many LLM calls only need a choice, like spam or not spam. Typed decision models answer these in a single pass, which fails when a decision needs several dependent steps. SanSi loops Ouro-1.4B, a model pretrained to reuse its layers, and reads option probabilities after each of 8 loops. Every loop is trained to answer, so the same model can stop anywhere from 1 to 8 loops. On 10,027 test decisions, SanSi scored 72.0% versus 58.4% for a same-shape model without loops, at 7.7x the compute. On liar chains longer than any seen in training, it reached 72.6% while a 4.2B single-pass model scored 50.0%, the chance level. If memory limits your decision-style calls, try a looped small model before moving to a model 3x larger.1h
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    1 Source

    Rohan Paul@rohanpaul_aiThis paper finds that small decision models get far more accurate by rerunning the same layers several times before giving an answer. Many LLM calls only need a choice, like spam or not spam. Typed decision models answer these in a single pass, which fails when a decision needs several dependent steps. SanSi loops Ouro-1.4B, a model pretrained to reuse its layers, and reads option probabilities after each of 8 loops. Every loop is trained to answer, so the same model can stop anywhere from 1 to 8 loops. On 10,027 test decisions, SanSi scored 72.0% versus 58.4% for a same-shape model without loops, at 7.7x the compute. On liar chains longer than any seen in training, it reached 72.6% while a 4.2B single-pass model scored 50.0%, the chance level. If memory limits your decision-style calls, try a looped small model before moving to a model 3x larger.1h
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet