© 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. 15092
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "com" related to URLs

    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
    397
    +0.19
    1.1%
    78
    +0.14
    0.8%
    34
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    500
    +0.19
    0.02
    34
    +0.14
    0.03
    482
    +0.13
    0.01
    Negative Logits
     suit
    -1.63
     precisely
    -1.57
     concerts
    -1.50
     exhibit
    -1.50
    .’”
    -1.37
     testament
    -1.34
     Parenthood
    -1.30
     govern
    -1.29
     approximate
    -1.28
     medicine
    -1.27
    POSITIVE LOGITS
    °
    4.20
    ¹
    4.08
    ¾
    3.95
    ¬
    3.88
    ´
    3.83
    ĭ
    3.82
    Į
    3.77
    ¶
    3.73
    ĵ
    3.73
    ·
    3.72
    Activations Density 0.046%

    No Known Activations