© 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. 12439
    Prev
    Next
    INDEX
    Explanations

    commands or requests related to sending or exchanging information

    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
    125
    +0.16
    0.9%
    53
    +0.14
    0.8%
    150
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    125
    +0.16
    0.05
    457
    +0.14
    0.04
    150
    +0.12
    0.04
    Negative Logits
    »
    -1.95
    ı
    -1.95
    ĸ´
    -1.94
    ļ
    -1.93
    ¶
    -1.89
    ½
    -1.85
    "}](#
    -1.84
    ¹
    -1.83
    Į
    -1.78
    Ĥ¬
    -1.75
    POSITIVE LOGITS
    imental
    2.99
    inel
    2.83
     messages
    2.78
     mail
    2.45
     message
    2.36
    ient
    2.32
     signals
    2.23
     emails
    2.10
     out
    2.10
    mail
    2.10
    Activations Density 0.483%

    No Known Activations