© 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. 26197
    Prev
    Next
    INDEX
    Explanations

    mathematical terms and concepts related to derivatives and factors

    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
    347
    +0.14
    0.8%
    145
    +0.12
    0.7%
    109
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    388
    +0.14
    0.02
    171
    +0.12
    0.01
    145
    +0.12
    0.03
    Negative Logits
     magn
    -1.39
    lections
    -1.37
    ographics
    -1.30
    uth
    -1.26
     %
    -1.24
    itively
    -1.24
    lessly
    -1.24
    >&
    -1.23
    rapeutics
    -1.23
    phrine
    -1.22
    POSITIVE LOGITS
    ener
    1.54
    bour
    1.54
    ucker
    1.46
    angel
    1.38
     maneuver
    1.36
    bian
    1.32
    PI
    1.29
    è»
    1.29
    cous
    1.29
    kap
    1.29
    Activations Density 0.289%

    No Known Activations