© 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. 4-RES-SM
    11. 14722
    Prev
    Next
    INDEX
    Explanations

    code-related constructs and chemical processes involving cells and fuel cells

    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
    274
    +0.14
    0.7%
    440
    +0.13
    0.7%
    183
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    47
    +0.14
    0.03
    399
    +0.13
    -0.01
    397
    +0.11
    0.02
    Negative Logits
     audit
    -1.60
     Abdullah
    -1.57
    shake
    -1.56
     Christie
    -1.56
    Prince
    -1.51
    Certificate
    -1.51
     Ross
    -1.50
    aud
    -1.48
     ski
    -1.46
     aud
    -1.43
    POSITIVE LOGITS
     cell
    4.58
     cells
    4.27
     Cell
    4.10
     Cells
    3.78
    cell
    3.74
    Cell
    3.65
    cells
    3.55
    Cells
    3.25
     cellular
    2.90
     Cellular
    2.90
    Activations Density 0.633%

    No Known Activations