© 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 EvalsExportsAPI 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. 27917
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "that" with varying contextual meanings

    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
    175
    +0.12
    0.7%
    23
    +0.11
    0.6%
    46
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    175
    +0.12
    0.04
    433
    +0.11
    0.04
    478
    +0.11
    0.00
    Negative Logits
     ourselves
    -1.68
    ums
    -1.60
     voice
    -1.50
    endar
    -1.44
    áĢ¸
    -1.38
     Duty
    -1.34
    feit
    -1.31
     trailers
    -1.30
     taps
    -1.29
     Fleming
    -1.29
    POSITIVE LOGITS
     slightest
    1.83
     ')
    1.80
     II
    1.57
    forth
    1.56
    xhtml
    1.45
     '.
    1.44
     ',
    1.42
    ).$
    1.41
    ),$$
    1.39
    OOGLE
    1.38
    Activations Density 0.653%

    No Known Activations