© 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. 13420
    Prev
    Next
    INDEX
    Explanations

    sequences or formats of numerical data and their respective 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
    369
    +0.13
    0.7%
    475
    +0.11
    0.6%
    483
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    369
    +0.13
    0.02
    483
    +0.11
    0.05
    475
    +0.11
    0.04
    Negative Logits
    comes
    -1.89
    end
    -1.62
    ensing
    -1.56
    ployment
    -1.52
    haven
    -1.50
    itance
    -1.45
     wherever
    -1.44
    iting
    -1.44
    lectual
    -1.41
    off
    -1.40
    POSITIVE LOGITS
    rd
    4.19
    Ŀ
    3.61
    Į
    3.23
    ļ
    3.20
    Ĵ
    3.18
    ´
    3.08
    Ļ
    3.08
    į
    3.03
    Ķ
    2.95
    İ
    2.91
    Activations Density 0.210%

    No Known Activations