© 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. MLP Post
    9. 0-MLP-SM
    11. 31182
    Prev
    Next
    INDEX
    Explanations

    references to the Catholic Church and its teachings

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__mlp-sm_processed/0-mlp-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.0.hook_mlp_out
    Hook Layer
    0
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    71
    +0.17
    1.0%
    156
    +0.16
    0.9%
    433
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    71
    +0.17
    0.02
    433
    +0.16
    0.02
    159
    +0.13
    0.02
    Negative Logits
    ľ
    -1.99
     distant
    -1.81
     varying
    -1.69
     different
    -1.60
    ĥ½
    -1.59
     crashing
    -1.58
     severed
    -1.56
     violent
    -1.50
     swollen
    -1.47
     pertinent
    -1.41
    POSITIVE LOGITS
    ism
    1.83
    nikov
    1.78
    mium
    1.72
    istan
    1.61
    ictionary
    1.61
    8859
    1.60
    opan
    1.59
    icip
    1.58
    BPF
    1.57
    FC
    1.54
    Activations Density 0.024%

    No Known Activations