© 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. 27967
    Prev
    Next
    INDEX
    Explanations

    punctuation marks or separators in text

    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
    47
    +0.13
    0.7%
    478
    +0.13
    0.7%
    368
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    157
    +0.13
    0.06
    13
    +0.13
    0.04
    47
    +0.12
    0.05
    Negative Logits
    kered
    -1.86
    dered
    -1.84
    zed
    -1.80
    hered
    -1.80
    falls
    -1.80
    ided
    -1.76
    fall
    -1.72
    imated
    -1.69
    .](
    -1.66
    xygen
    -1.62
    POSITIVE LOGITS
    ´
    3.55
    ¥
    3.50
    IJ
    3.35
    į
    3.34
    Īĺ
    3.34
    ī
    3.30
    ®
    3.27
    ¬
    3.22
    µ
    3.22
    ľ
    3.14
    Activations Density 0.272%

    No Known Activations