© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 23886
    Prev
    Next
    INDEX
    Explanations

    infinitive verbs indicating actions or processes

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    23
    +0.21
    1.2%
    411
    +0.13
    0.8%
    19
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    411
    +0.21
    0.06
    19
    +0.13
    0.04
    151
    +0.13
    0.04
    Negative Logits
    ĩ
    -3.07
    Ĭ
    -2.88
    ¡
    -2.86
    »¿
    -2.78
    ¢
    -2.78
    Ģ
    -2.77
    Ķ
    -2.77
    ĻĤ
    -2.74
    ŀ
    -2.69
    ↵  ³³³
    -2.68
    POSITIVE LOGITS
    oes
    2.10
     handle
    1.97
    yx
    1.78
    opy
    1.77
     successfully
    1.70
    bey
    1.66
     cope
    1.65
    acity
    1.64
     Redistributions
    1.59
    +$
    1.56
    Activations Density 0.262%

    No Known Activations