© 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. 2-RES-SM
    11. 27630
    Prev
    Next
    INDEX
    Explanations

    words indicating a connection or conjunction between ideas or elements

    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/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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    95
    +0.13
    0.7%
    288
    +0.12
    0.6%
    153
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    95
    +0.13
    0.06
    335
    +0.12
    0.06
    266
    +0.11
    0.05
    Negative Logits
    ientes
    -1.68
    ";
    -1.56
    eties
    -1.45
     boundaries
    -1.39
     acceptance
    -1.38
    ificates
    -1.36
    ubes
    -1.36
     shares
    -1.35
     Ocean
    -1.33
     assurance
    -1.32
    POSITIVE LOGITS
    áĢ¯
    2.04
    áĢº
    1.96
    à°¿
    1.87
    à³
    1.78
    áĢ¬
    1.77
    bid
    1.71
    erd
    1.69
    à±į
    1.67
    iginally
    1.67
    à±ģ
    1.64
    Activations Density 0.610%

    No Known Activations