hypothesis
active
hypothesis:simplicity-bias-hypothesis

Simplicity Bias Hypothesis

Deep networks are biased toward finding simple fits to data, and this bias increases with model size, driving convergence

Source paper

extracted_from
The Platonic Representation Hypothesis
(2024) · Minyoung Huh · Brian Cheung · Tongzhou Wang · Phillip Isola

Neighborhood — ranked by edge-count

Concepts (1)

concept
  • The hypothesized converged representation that all sufficiently large AI models are approaching — a statistical model of underlying reality

Hypotheses (1)

hypothesis
  • Capacity Hypothesis
    associated_with
    Bigger models are more likely to converge to a shared representation than smaller models because they can better approximate the global optimum

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

  • Simplicity Biasconcept0.883
    The tendency of deep networks to implicitly favor simpler solutions that fit the data, driving convergence
  • Reward Hypothesisconcept0.772
    The claim in RL that any goal can be expressed as maximizing the expected cumulative sum of a scalar reward signal.
  • Inductive Biasconcept0.757
    Assumptions or preferences (e.g., parsimony) that determine how a learning system generalizes beyond training data
  • Features related to gender, racial, ethnic biases, slurs, and hate speech.
  • Ability to entertain competing hypotheses within one inference engine; proposed hallmark of mindful inference
  • Genesis Hypothesisframework0.750
    The conjecture that consciousness does not result from the organized mind but creates and maintains complex models of reality; forms at the beginning of mental development
  • The hypothesis that analogous features and circuits reliably form across different neural network models and tasks
  • Bias Amplificationconcept0.744
    Problem cited as a limitation of current LLMs; PRH predicts larger models should amplify bias less