© 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. 4-RES-SM
    11. 15223
    Prev
    Next
    INDEX
    Explanations

    demographic data and statistics related to communities

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    21
    +0.16
    0.9%
    78
    +0.13
    0.8%
    314
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    30
    +0.16
    0.01
    203
    +0.13
    0.02
    314
    +0.12
    0.03
    Negative Logits
    actic
    -1.47
     helper
    -1.46
     repertoire
    -1.40
     priming
    -1.39
     overload
    -1.38
     batches
    -1.38
     cassette
    -1.38
     incompatible
    -1.37
     soon
    -1.37
    !’
    -1.34
    POSITIVE LOGITS
     longitude
    1.76
     land
    1.72
     borders
    1.70
     map
    1.67
    iterranean
    1.62
     mountains
    1.61
     municipality
    1.55
     median
    1.54
     Census
    1.53
     census
    1.53
    Activations Density 0.158%

    No Known Activations