© 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. 2-RES-SM
    11. 9959
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "constant" and variations thereof

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    488
    +0.17
    0.9%
    217
    +0.15
    0.8%
    343
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    488
    +0.17
    0.01
    343
    +0.15
    0.02
    29
    +0.12
    0.01
    Negative Logits
    Ķ
    -1.80
    oard
    -1.61
    orent
    -1.57
    kers
    -1.56
    inth
    -1.54
    cons
    -1.54
    acco
    -1.54
    them
    -1.53
    ª
    -1.52
    mit
    -1.48
    POSITIVE LOGITS
    ness
    1.72
    ely
    1.70
    itudes
    1.57
    ially
    1.54
    ise
    1.54
     reference
    1.50
    ierung
    1.50
    iosity
    1.48
    izability
    1.48
    ive
    1.48
    Activations Density 0.094%

    No Known Activations