© 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. 2-RES-SM
    11. 14305
    Prev
    Next
    INDEX
    Explanations

    numerical data and sorting terms related to order

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    151
    +0.13
    0.8%
    102
    +0.11
    0.6%
    333
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    322
    +0.13
    0.01
    82
    +0.11
    0.02
    125
    +0.11
    0.02
    Negative Logits
     customs
    -1.72
    edly
    -1.64
    onom
    -1.60
    onia
    -1.59
    thon
    -1.48
    oth
    -1.45
     inheritance
    -1.42
     named
    -1.42
    rim
    -1.40
    ounty
    -1.39
    POSITIVE LOGITS
    enne
    1.60
    zza
    1.49
    EW
    1.43
    Ú
    1.42
    DTD
    1.35
    fefe
    1.35
    IMO
    1.32
    FE
    1.30
    iator
    1.28
    entre
    1.27
    Activations Density 0.054%

    No Known Activations