© 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. 2-RES-SM
    11. 27637
    Prev
    Next
    INDEX
    Explanations

    instances of parentheses

    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
    156
    +0.21
    1.2%
    478
    +0.20
    1.2%
    188
    +0.16
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    188
    +0.21
    0.06
    350
    +0.20
    0.06
    219
    +0.16
    0.05
    Negative Logits
    opan
    -1.69
    indows
    -1.56
    ervice
    -1.53
    ween
    -1.53
    "(
    -1.50
    ios
    -1.50
    inical
    -1.47
    rically
    -1.45
    """
    -1.44
    eful
    -1.44
    POSITIVE LOGITS
    Īĺ
    3.25
    ĻĤ
    3.11
    ´
    3.04
    ¦
    3.02
    Ľ
    2.94
    Ķ
    2.85
    IJ
    2.83
    ĺ
    2.79
    Ĭ
    2.75
    «
    2.69
    Activations Density 0.201%

    No Known Activations