© 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. 1-RES-SM
    11. 26087
    Prev
    Next
    INDEX
    Explanations

    instances of various forms of the verb "to be" indicating presence or existence

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    23
    +0.40
    2.4%
    478
    +0.13
    0.7%
    348
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    348
    +0.40
    0.06
    485
    +0.13
    0.04
    23
    +0.12
    0.05
    Negative Logits
    facts
    -1.86
    forum
    -1.83
    works
    -1.83
    mind
    -1.80
    pieces
    -1.78
    quarters
    -1.76
    lette
    -1.73
    door
    -1.71
     minds
    -1.68
    haus
    -1.66
    POSITIVE LOGITS
     pronounced
    2.10
    poons
    2.10
    ]{.
    1.91
     evenly
    1.73
     supposed
    1.68
     symmetric
    1.65
     trypsin
    1.64
    hers
    1.59
     agarose
    1.58
     probable
    1.56
    Activations Density 0.184%

    No Known Activations