© 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. 25606
    Prev
    Next
    INDEX
    Explanations

    mentions of the name "Thomas."

    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
    156
    +0.17
    1.0%
    369
    +0.14
    0.8%
    352
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    385
    +0.17
    0.03
    352
    +0.14
    0.03
    71
    +0.13
    0.02
    Negative Logits
    ĥ½
    -2.13
    ĻĤ
    -2.09
    uties
    -1.95
    ·¸
    -1.81
    ¥
    -1.76
    ı
    -1.73
    ordinate
    -1.65
    ĥ
    -1.63
    ¬
    -1.63
    ¼
    -1.60
    POSITIVE LOGITS
    ville
    2.20
     Edison
    2.04
    quito
    2.01
    chitz
    1.98
    boro
    1.84
    stown
    1.82
    pora
    1.80
    iewicz
    1.78
    bilt
    1.74
    creen
    1.63
    Activations Density 0.079%

    No Known Activations