© 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. 27946
    Prev
    Next
    INDEX
    Explanations

    instances of the word "shown" related to scientific findings or observations

    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
    223
    +0.12
    0.7%
    368
    +0.12
    0.7%
    277
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    115
    +0.12
    0.01
    372
    +0.12
    0.01
    283
    +0.11
    0.02
    Negative Logits
    »¿
    -1.79
    Ń
    -1.76
    Ł
    -1.69
    ¡
    -1.69
    owned
    -1.62
    robe
    -1.60
    dule
    -1.58
    ĩ
    -1.58
    IJ
    -1.57
    ī
    -1.56
    POSITIVE LOGITS
     prominently
    1.84
     respect
    1.82
     glimps
    1.81
     examples
    1.76
     how
    1.71
     similarities
    1.66
    eking
    1.65
     clearly
    1.63
    ishes
    1.58
     regard
    1.58
    Activations Density 0.073%

    No Known Activations