© 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. 9680
    Prev
    Next
    INDEX
    Explanations

    alphanumeric sequences and special characters in technical or data-related contexts

    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
    198
    +0.19
    1.1%
    216
    +0.13
    0.8%
    235
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.19
    0.06
    17
    +0.13
    0.01
    26
    +0.13
    0.04
    Negative Logits
    Ļª
    -2.37
    ĭ
    -2.02
    Ĺ
    -1.92
    unes
    -1.72
    à±į
    -1.72
    à±ģ
    -1.68
    àµį
    -1.67
    Ļ
    -1.65
    Ń
    -1.63
    į
    -1.62
    POSITIVE LOGITS
    ickets
    1.47
     admission
    1.43
     Zn
    1.41
    eness
    1.41
     Heavy
    1.39
     hops
    1.39
     Boys
    1.34
    oreal
    1.29
     Box
    1.29
     firmly
    1.28
    Activations Density 0.998%

    No Known Activations