© 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 EvalsExportsAPI 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. 11273
    Prev
    Next
    INDEX
    Explanations

    clauses that outline legal conditions or regulations

    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
    +0.15
    0.8%
    271
    +0.12
    0.7%
    318
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    203
    +0.15
    0.03
    222
    +0.12
    0.02
    30
    +0.10
    0.00
    Negative Logits
    leton
    -1.54
    rene
    -1.46
    led
    -1.45
    yon
    -1.36
     **
    -1.27
    FROM
    -1.27
    ?**
    -1.26
    edo
    -1.25
    assic
    -1.24
    uded
    -1.21
    POSITIVE LOGITS
    arta
    1.66
     WARRANTY
    1.46
    )](
    1.46
    )$)
    1.46
    dz
    1.43
     paste
    1.38
     drums
    1.37
    amsfonts
    1.31
    oltz
    1.29
    amssymb
    1.26
    Activations Density 0.633%

    No Known Activations