© 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. 14153
    Prev
    Next
    INDEX
    Explanations

    hexadecimal numerical values and strings related to coding or software functions

    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
    78
    +0.20
    1.1%
    348
    +0.14
    0.8%
    478
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    136
    +0.20
    0.00
    502
    +0.14
    0.06
    151
    +0.11
    0.04
    Negative Logits
     standpoint
    -1.80
     viewpoint
    -1.65
     neutrality
    -1.60
     perspective
    -1.53
    manship
    -1.47
     rather
    -1.41
     only
    -1.41
    pective
    -1.39
     soon
    -1.38
    osecond
    -1.38
    POSITIVE LOGITS
    1451450014514500
    2.61
    aeda
    2.31
    eba
    2.01
    cb
    1.99
    bd
    1.93
    dbc
    1.93
    fefe
    1.87
    ){#
    1.84
    ea
    1.84
    iox
    1.83
    Activations Density 0.218%

    No Known Activations