© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 13778
    Prev
    Next
    INDEX
    Explanations

    phrases related to components and specifications in technical or engineering contexts

    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/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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    25
    +0.27
    1.7%
    307
    +0.23
    1.5%
    466
    +0.14
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    307
    +0.27
    0.09
    296
    +0.23
    0.08
    282
    +0.14
    0.06
    Negative Logits
    ienn
    -1.54
    \][
    -1.49
     =>
    -1.43
    ermo
    -1.43
    brew
    -1.38
    oster
    -1.38
    iji
    -1.35
    *](#
    -1.33
    rud
    -1.28
    matic
    -1.27
    POSITIVE LOGITS
     each
    2.47
    ĻĤ
    2.43
     these
    2.34
    ĥ½
    2.33
    fered
    2.32
     interest
    2.17
    ¿½
    2.10
     individual
    1.95
    Ļª
    1.93
     the
    1.89
    Activations Density 0.726%

    No Known Activations