© 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. 13648
    Prev
    Next
    INDEX
    Explanations

    code structures related to class definitions and method declarations in programming 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
    21
    +0.21
    1.2%
    78
    +0.21
    1.2%
    388
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    22
    +0.21
    0.04
    185
    +0.21
    0.02
    78
    +0.13
    0.00
    Negative Logits
    ising
    -1.45
    izing
    -1.38
    umer
    -1.32
     guides
    -1.32
    ulating
    -1.31
     indeed
    -1.30
     dice
    -1.28
     odds
    -1.25
     pla
    -1.25
     hub
    -1.23
    POSITIVE LOGITS
    {
    2.45
    {};
    1.97
    Ģ
    1.93
    Ĭ
    1.90
    §
    1.86
    {}
    1.78
    inherit
    1.74
    à±ģ
    1.74
    à±
    1.74
    ģ
    1.72
    Activations Density 0.156%

    No Known Activations