© 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
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 4-RES-SM
    6. 9333
    Prev
    Next
    INDEX
    Explanations

    elements of artistic expression

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    +0.20
    1.1%
    307
    +0.14
    0.8%
    323
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.20
    0.09
    71
    +0.14
    0.08
    274
    +0.12
    0.09
    Negative Logits
    "}](#
    -2.07
    \]\].
    -1.86
    amycin
    -1.69
     mailing
    -1.62
    ="$
    -1.56
     ones
    -1.54
    asone
    -1.51
    ="../../../../../../
    -1.49
     purposes
    -1.46
     wonders
    -1.42
    POSITIVE LOGITS
     Its
    3.25
     It
    2.65
     Originally
    2.57
     Initially
    2.53
     This
    2.53
     Their
    2.53
     Unlike
    2.50
     They
    2.31
     Although
    2.30
     Each
    2.30
    Activations Density 0.506%

    No Known Activations