© 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. 1-RES-SM
    11. 29728
    Prev
    Next
    INDEX
    Explanations

    references to programming-related concepts and syntax, particularly around regex and its usage

    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/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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    289
    +0.17
    1.0%
    439
    +0.13
    0.7%
    4
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    157
    +0.17
    0.05
    17
    +0.13
    0.01
    274
    +0.12
    0.04
    Negative Logits
    ³
    -2.19
    Ĺ
    -2.14
    Ŀ
    -2.06
    ļ
    -1.97
    ŀ
    -1.97
    ¯
    -1.90
    ¨
    -1.87
    Ĥ
    -1.87
    £
    -1.83
    §
    -1.82
    POSITIVE LOGITS
     regex
    1.88
     keywords
    1.85
    regex
    1.76
    slash
    1.73
    fasterxml
    1.71
     tags
    1.69
     comma
    1.69
     tabs
    1.69
     chars
    1.69
     spam
    1.67
    Activations Density 0.541%

    No Known Activations