© 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. 28890
    Prev
    Next
    INDEX
    Explanations

    sections of code with function declarations and their corresponding parameters

    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
    210
    +0.11
    0.6%
    509
    +0.11
    0.6%
    366
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    53
    +0.11
    0.11
    366
    +0.11
    0.07
    461
    +0.10
    0.07
    Negative Logits
    áĢ¯
    -1.62
     Bulletin
    -1.40
    Series
    -1.37
     Dogs
    -1.37
     Risk
    -1.35
     Encyclopedia
    -1.34
     Chronicle
    -1.33
    áĢ¬
    -1.32
    Stock
    -1.30
     Heat
    -1.29
    POSITIVE LOGITS
    gle
    1.56
    merce
    1.52
    =>
    1.50
    eu
    1.42
     =>
    1.41
     upward
    1.40
    ```
    1.39
     unequal
    1.37
     here
    1.37
    ów
    1.35
    Activations Density 1.028%

    No Known Activations