© 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. 8840
    Prev
    Next
    INDEX
    Explanations

    modal verbs indicating ability or possibility

    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
    235
    +0.11
    0.6%
    380
    +0.11
    0.6%
    234
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    435
    +0.11
    0.07
    329
    +0.11
    0.06
    272
    +0.10
    0.05
    Negative Logits
    ி
    -2.02
    à¯ģ
    -1.81
    à¯
    -1.78
    ató
    -1.74
    ivism
    -1.73
    rays
    -1.68
    icos
    -1.64
    hea
    -1.63
    imony
    -1.58
    cology
    -1.58
    POSITIVE LOGITS
    't
    2.56
     choose
    2.37
     easily
    2.29
     customize
    2.17
     see
    2.07
     also
    2.01
     write
    1.98
     always
    1.96
     access
    1.95
     swap
    1.87
    Activations Density 0.289%

    No Known Activations