© 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. 24298
    Prev
    Next
    INDEX
    Explanations

    the word "are" in various contexts, indicating a focus on existence or quantity

    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
    23
    +0.30
    1.8%
    478
    +0.17
    1.0%
    442
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    235
    +0.30
    0.05
    442
    +0.17
    0.05
    383
    +0.12
    0.04
    Negative Logits
    ł
    -4.43
    §
    -4.18
    »¿
    -4.03
    İ
    -4.01
    ĨĴ
    -3.80
    ©
    -3.79
    Ļ
    -3.79
    ī
    -3.68
    ¡
    -3.64
    ¥
    -3.60
    POSITIVE LOGITS
     hereby
    1.65
     gonna
    1.59
     also
    1.58
     two
    1.47
    zzo
    1.47
    itions
    1.44
     supposed
    1.43
    irs
    1.43
    wolf
    1.42
    uka
    1.42
    Activations Density 0.241%

    No Known Activations