© 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. 10775
    Prev
    Next
    INDEX
    Explanations

    code structures or syntax elements within programming or markup languages

    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
    +0.23
    1.3%
    9
    +0.21
    1.2%
    466
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    9
    +0.23
    0.05
    78
    +0.21
    0.01
    502
    +0.11
    0.06
    Negative Logits
    doms
    -1.60
     discretion
    -1.43
    ftware
    -1.40
    opathy
    -1.37
    bucks
    -1.36
    oxford
    -1.35
    balls
    -1.35
    abouts
    -1.34
    ball
    -1.33
     deposit
    -1.33
    POSITIVE LOGITS
    ·
    2.81
    «
    2.78
    ł
    2.69
    ĭ
    2.55
    ķ
    2.42
    Ĭ
    2.36
    Ļª
    2.34
    ®
    2.33
    Į
    2.27
    ¦
    2.25
    Activations Density 0.346%

    No Known Activations