© 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
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 4-RES-SM
    6. 9702
    Prev
    Next
    INDEX
    Explanations

    references to organizational structures and entities involved in governance

    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
    +0.36
    2.2%
    279
    +0.13
    0.8%
    35
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    21
    +0.36
    0.05
    234
    +0.13
    0.05
    186
    +0.11
    0.06
    Negative Logits
     of
    -2.76
    Of
    -2.20
     Of
    -2.20
    of
    -1.88
     OF
    -1.87
     thereof
    -1.71
     dei
    -1.56
    ococcus
    -1.55
    ·¸
    -1.53
     eines
    -1.40
    POSITIVE LOGITS
    ÑģÑĤва
    1.51
    xico
    1.39
    matically
    1.38
    edly
    1.29
    heres
    1.29
    cludes
    1.28
    sburg
    1.26
     xenograft
    1.25
    cedes
    1.25
    arily
    1.20
    Activations Density 1.889%

    No Known Activations