© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 8742
    Prev
    Next
    INDEX
    Explanations

    phrases indicating people's opinions or beliefs

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/4-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    377
    +0.14
    0.8%
    86
    +0.13
    0.7%
    420
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    420
    +0.14
    0.12
    406
    +0.13
    0.10
    320
    +0.12
    0.15
    Negative Logits
    ĻĤ
    -5.00
    ĥ½
    -4.64
    ¯
    -4.60
    Ħ
    -4.51
    ĨĴ
    -4.38
    ij
    -4.37
      ↵
    -4.32
     ↵
    -4.32
    ↵
    -4.32
    <|outofrange|>
    -4.32
    POSITIVE LOGITS
     wondering
    2.73
     wonder
    2.64
     wondered
    2.56
     think
    2.52
     envy
    2.47
     saliv
    2.45
     believe
    2.43
     speculate
    2.41
     doubt
    2.37
     admire
    2.35
    Activations Density 1.930%

    No Known Activations