© 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. 31918
    Prev
    Next
    INDEX
    Explanations

    expressions of surprise or laughter

    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
    87
    +0.23
    1.3%
    241
    +0.18
    1.1%
    450
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.23
    0.02
    241
    +0.18
    0.03
    410
    +0.14
    0.03
    Negative Logits
    ¿½
    -2.02
    ¬
    -1.92
    verted
    -1.81
    ĨĴ
    -1.76
    ellular
    -1.71
    ij
    -1.70
    ¢
    -1.66
    ODING
    -1.65
    izable
    -1.61
    clud
    -1.57
    POSITIVE LOGITS
    olics
    1.76
    noon
    1.75
     Abdul
    1.60
    ines
    1.54
    orns
    1.53
    ulu
    1.51
    iri
    1.49
    mina
    1.47
    ouses
    1.46
    aron
    1.45
    Activations Density 0.111%

    No Known Activations