© 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. 13619
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "make" in various forms

    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
    307
    +0.25
    1.4%
    447
    +0.18
    1.1%
    286
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    447
    +0.25
    0.07
    303
    +0.18
    0.06
    286
    +0.13
    0.07
    Negative Logits
    ij
    -1.61
    «
    -1.56
    ancy
    -1.56
    İ
    -1.55
    ļ
    -1.51
    yi
    -1.48
    Ĥ¬
    -1.46
    itus
    -1.46
    gium
    -1.46
    encies
    -1.44
    POSITIVE LOGITS
     sure
    3.07
    hift
    2.62
    leine
    2.17
     sense
    2.00
     use
    1.93
    sure
    1.75
     adjustments
    1.75
     clear
    1.69
     Sense
    1.67
     haste
    1.67
    Activations Density 0.129%

    No Known Activations