© 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. 15323
    Prev
    Next
    INDEX
    Explanations

    phrases indicating questions or advice on how to do something

    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/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
    Neuron Alignment
    Index
    Value
    % of L₁
    231
    +0.14
    0.8%
    358
    +0.12
    0.7%
    94
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    94
    +0.14
    0.04
    231
    +0.12
    0.04
    358
    +0.11
    0.05
    Negative Logits
    »¿
    -2.29
    Ļª
    -1.78
    laughter
    -1.63
    )$)
    -1.60
    oslav
    -1.59
    ி
    -1.58
     itself
    -1.57
    ãĤī
    -1.56
    ipart
    -1.56
    \]]{}
    -1.53
    POSITIVE LOGITS
     handle
    2.43
     avoid
    2.33
     distribute
    2.24
     circumvent
    2.23
     configure
    2.23
     calculate
    2.20
     achieve
    2.18
     reconcile
    2.17
     dispose
    2.15
     minimize
    2.15
    Activations Density 0.226%

    No Known Activations