© 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. 9176
    Prev
    Next
    INDEX
    Explanations

    phrases and actions related to writing and documentation

    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
    307
    +0.16
    0.9%
    457
    +0.13
    0.7%
    230
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    457
    +0.16
    0.03
    55
    +0.13
    0.04
    398
    +0.12
    0.01
    Negative Logits
    «
    -3.00
    ·
    -2.71
    ¨
    -2.65
    ļ
    -2.62
    ı
    -2.62
    ª
    -2.49
    ¹
    -2.47
    ĸ
    -2.42
    ij
    -2.41
    Ł
    -2.35
    POSITIVE LOGITS
     about
    2.44
     poetry
    2.09
     down
    2.04
     fiction
    1.93
     scripts
    1.91
     poems
    1.87
     code
    1.80
     something
    1.70
    pone
    1.70
     novels
    1.67
    Activations Density 0.316%

    No Known Activations