© 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. 27923
    Prev
    Next
    INDEX
    Explanations

    keywords related to authentication and login processes

    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
    400
    +0.16
    0.9%
    156
    +0.15
    0.9%
    497
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    340
    +0.16
    0.06
    400
    +0.15
    0.04
    497
    +0.14
    0.04
    Negative Logits
    ophagus
    -1.84
    tera
    -1.65
    opard
    -1.54
     purse
    -1.49
     culprit
    -1.46
     place
    -1.46
    clud
    -1.33
    creas
    -1.32
    iscus
    -1.31
    IGNED
    -1.30
    POSITIVE LOGITS
    Ļª
    2.17
    azine
    1.85
    ©
    1.64
    ļ
    1.58
    ĨĴ
    1.57
    ĸ
    1.50
    ŀ
    1.48
    ³
    1.45
    .
    1.44
     faithfully
    1.43
    Activations Density 1.302%

    No Known Activations