© 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. 2-RES-SM
    11. 14143
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, particularly those indicating dialogue and excitement

    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/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
    Neuron Alignment
    Index
    Value
    % of L₁
    23
    +0.15
    0.9%
    17
    +0.12
    0.6%
    290
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    103
    +0.15
    0.06
    133
    +0.12
    0.05
    23
    +0.11
    0.07
    Negative Logits
    ól
    -1.74
     chemistry
    -1.56
    heets
    -1.48
     known
    -1.47
    ruptcy
    -1.46
     participating
    -1.44
    ario
    -1.41
    hold
    -1.39
    urities
    -1.35
     elsewhere
    -1.35
    POSITIVE LOGITS
    į
    3.13
    ³
    2.83
    ī
    2.71
    ¼
    2.69
    Ĩ
    2.66
    ¬
    2.54
    ¹
    2.53
    ¢
    2.50
    ĩ
    2.46
    Ĺ
    2.46
    Activations Density 0.499%

    No Known Activations