© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 30138
    Prev
    Next
    INDEX
    Explanations

    technical specifications related to devices and materials

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    23
    +0.27
    1.5%
    67
    +0.13
    0.7%
    485
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.27
    0.08
    8
    +0.13
    0.05
    83
    +0.11
    0.05
    Negative Logits
    ERC
    -2.02
    İ
    -1.98
    ¡
    -1.98
    ł
    -1.97
    ¥
    -1.96
     expires
    -1.89
    ĵ
    -1.84
    ¢
    -1.83
    ĸ
    -1.76
    į
    -1.71
    POSITIVE LOGITS
     desir
    1.66
    orously
    1.65
     xe
    1.64
    rapeut
    1.56
     particularly
    1.56
    hes
    1.55
    tol
    1.54
     hereby
    1.51
     separation
    1.51
     optionally
    1.51
    Activations Density 0.452%

    No Known Activations