© 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. 1-RES-SM
    11. 29969
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "push" and variations related to pushing actions

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/1-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.1.hook_resid_post
    Hook Layer
    1
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    59
    +0.17
    1.0%
    452
    +0.14
    0.8%
    317
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    452
    +0.17
    0.03
    59
    +0.14
    0.03
    317
    +0.13
    0.02
    Negative Logits
     trusted
    -1.46
     slightest
    -1.40
     Shepherd
    -1.38
     she
    -1.34
    documentclass
    -1.34
     favour
    -1.34
     convicted
    -1.27
     scent
    -1.26
     favor
    -1.26
     try
    -1.25
    POSITIVE LOGITS
    į
    2.19
    ¬
    2.18
    Ł
    2.16
    ĸ
    2.16
    ĩ
    2.12
    Ħ
    2.11
    ī
    2.06
    ¡
    2.05
    IJ
    2.01
    Ģ
    2.01
    Activations Density 0.115%

    No Known Activations