© 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
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 14005
    Prev
    Next
    INDEX
    Explanations

    terms related to regular expressions (regex) and their components

    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
    307
    +0.14
    0.8%
    156
    +0.14
    0.8%
    77
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    193
    +0.14
    0.04
    334
    +0.14
    0.03
    9
    +0.13
    0.02
    Negative Logits
     campus
    -1.48
    organ
    -1.41
     taste
    -1.37
     amen
    -1.35
    irts
    -1.35
    ksi
    -1.34
     Cir
    -1.34
    irt
    -1.33
     ongoing
    -1.31
     campuses
    -1.31
    POSITIVE LOGITS
    aline
    2.25
    hematic
    2.18
    uximab
    2.11
    pot
    2.00
    finder
    1.92
    listing
    1.89
    fony
    1.87
    fiddle
    1.86
    p
    1.83
    olver
    1.81
    Activations Density 0.286%

    No Known Activations