© 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. 15082
    Prev
    Next
    INDEX
    Explanations

    elements related to data structures and programming logic

    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
    21
    +0.25
    1.5%
    78
    +0.14
    0.9%
    149
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    415
    +0.25
    0.07
    21
    +0.14
    0.05
    432
    +0.11
    0.07
    Negative Logits
     contribution
    -1.57
     minds
    -1.52
     volume
    -1.36
    ibility
    -1.30
     continuation
    -1.29
    aries
    -1.29
     Bulletin
    -1.25
    apine
    -1.23
     nature
    -1.22
     behaviour
    -1.18
    POSITIVE LOGITS
    Ģ
    3.72
    ĥ½
    3.63
    ĵ
    3.61
    ĨĴ
    3.55
    Ī
    3.45
    ı
    3.44
    Ħ
    3.40
    ĭ
    3.34
    ij
    3.32
    ķ
    3.31
    Activations Density 0.262%

    No Known Activations