© 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. 30105
    Prev
    Next
    INDEX
    Explanations

    references to software and technical jargon

    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
    490
    +0.13
    0.7%
    304
    +0.12
    0.7%
    410
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    185
    +0.13
    0.02
    134
    +0.12
    0.02
    386
    +0.11
    0.02
    Negative Logits
    uously
    -1.71
    ductive
    -1.67
    most
    -1.61
     caution
    -1.52
    ucle
    -1.51
     blame
    -1.48
    ologically
    -1.43
    came
    -1.41
    forth
    -1.39
     deepest
    -1.36
    POSITIVE LOGITS
    àµį
    2.47
    AIL
    2.42
    ERY
    2.42
    àµ
    2.02
    ENG
    1.91
    eries
    1.87
    ield
    1.82
    IOR
    1.75
    ailable
    1.73
    inery
    1.70
    Activations Density 0.074%

    No Known Activations