© 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. 23986
    Prev
    Next
    INDEX
    Explanations

    the frequent use of the word "the" and related articles, alongside contextually relevant terms

    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
    156
    +0.14
    0.8%
    23
    +0.13
    0.7%
    18
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    10
    +0.14
    0.05
    128
    +0.13
    0.06
    98
    +0.12
    -0.00
    Negative Logits
    ale
    -1.63
    eros
    -1.56
    Ĥ¬
    -1.54
    vent
    -1.54
    »¿
    -1.50
    ÅĽÄĩ
    -1.49
    14514500
    -1.47
    Ĭ
    -1.46
    verte
    -1.46
    "];
    -1.45
    POSITIVE LOGITS
     footsteps
    1.84
     indeed
    1.59
    )):
    1.55
    )$)
    1.53
     users
    1.52
    hentication
    1.43
     sorts
    1.41
     sudden
    1.41
    ods
    1.40
     GNU
    1.40
    Activations Density 0.647%

    No Known Activations