© 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. 29762
    Prev
    Next
    INDEX
    Explanations

    suffixes related to actions or processes

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    232
    +0.12
    0.7%
    509
    +0.11
    0.7%
    177
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    509
    +0.12
    0.03
    311
    +0.11
    0.02
    232
    +0.11
    0.02
    Negative Logits
     nothing
    -1.94
     little
    -1.88
     place
    -1.87
     shit
    -1.83
     naught
    -1.78
     birth
    -1.72
     anything
    -1.68
     neither
    -1.63
     living
    -1.59
     maternal
    -1.55
    POSITIVE LOGITS
    naire
    3.34
    nier
    2.05
    uclide
    1.98
    arios
    1.84
    nel
    1.84
    istas
    1.83
    folio
    1.82
    kowski
    1.81
    nement
    1.79
    etics
    1.76
    Activations Density 0.100%

    No Known Activations