© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 13652
    Prev
    Next
    INDEX
    Explanations

    hyperlinks or references to external resources

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    337
    +0.12
    0.7%
    255
    +0.11
    0.6%
    332
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    174
    +0.12
    0.02
    185
    +0.11
    0.01
    219
    +0.10
    0.02
    Negative Logits
    tics
    -1.71
     Territory
    -1.59
    manship
    -1.56
    ycin
    -1.56
    apine
    -1.52
    onies
    -1.50
    ism
    -1.44
     incumbent
    -1.41
    oys
    -1.40
     namely
    -1.39
    POSITIVE LOGITS
    ³
    4.79
    ĨĴ
    4.62
    ¸
    4.43
    ¡
    4.37
    İ
    4.31
    ¦
    4.27
    Ļª
    4.26
    ±
    4.20
    IJ
    4.16
    ı
    4.14
    Activations Density 0.074%

    No Known Activations