© 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. 2-RES-SM
    11. 14203
    Prev
    Next
    INDEX
    Explanations

    references to locations, events, and entities related to entertainment and community engagement

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    161
    +0.18
    1.1%
    293
    +0.17
    1.0%
    419
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    210
    +0.18
    0.04
    293
    +0.17
    0.12
    161
    +0.15
    0.06
    Negative Logits
    axel
    -1.39
     '{
    -1.39
    ouch
    -1.36
     nuisance
    -1.26
    "?>
    -1.26
     Santos
    -1.24
    juvant
    -1.24
    idal
    -1.23
     mistakes
    -1.22
     mistake
    -1.21
    POSITIVE LOGITS
    CLU
    1.50
    cases
    1.41
     else
    1.38
     Weinstein
    1.35
    offs
    1.34
     also
    1.33
     Congress
    1.29
    onds
    1.29
     Awards
    1.29
     Families
    1.28
    Activations Density 3.970%

    No Known Activations