© 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. 13221
    Prev
    Next
    INDEX
    Explanations

    various punctuation marks and their occurrences in the text

    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
    +0.17
    1.0%
    510
    +0.16
    1.0%
    274
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    400
    +0.17
    0.05
    258
    +0.16
    0.12
    53
    +0.15
    0.06
    Negative Logits
    ¹
    -2.49
    Ģ
    -2.44
    į
    -2.15
    Ī
    -2.15
    ,^[@
    -2.12
    Ĭ
    -2.10
    ¶
    -2.10
    ª
    -2.07
    ¸
    -2.06
    §
    -2.02
    POSITIVE LOGITS
     The
    2.88
    The
    2.53
     Each
    2.42
     Its
    2.27
     There
    2.20
     Everything
    2.15
     It
    1.97
     Those
    1.94
     Their
    1.93
    Its
    1.92
    Activations Density 1.428%

    No Known Activations