© 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. 24881
    Prev
    Next
    INDEX
    Explanations

    occurrences of the modal verb "can."

    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
    468
    +0.16
    0.9%
    171
    +0.13
    0.7%
    285
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    468
    +0.16
    0.07
    202
    +0.13
    0.05
    166
    +0.11
    0.05
    Negative Logits
     heels
    -1.77
     heel
    -1.66
    )).
    -1.59
    ureus
    -1.56
    orse
    -1.55
    innen
    -1.54
    !”
    -1.51
     positives
    -1.50
     chances
    -1.49
    ees
    -1.47
    POSITIVE LOGITS
    've
    1.78
    'd
    1.59
    'm
    1.50
    ieurs
    1.49
    va
    1.46
     include
    1.45
     ded
    1.39
    't
    1.37
    'll
    1.36
    PHP
    1.36
    Activations Density 0.411%

    No Known Activations