© 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. 4-RES-SM
    11. 8675
    Prev
    Next
    INDEX
    Explanations

    sentences conveying emotional or dramatic content related to violence or conflict

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    307
    +0.25
    1.5%
    156
    +0.20
    1.1%
    25
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    307
    +0.25
    0.15
    268
    +0.20
    0.15
    466
    +0.14
    0.13
    Negative Logits
    ģ
    -2.63
    §
    -2.54
    ŀ
    -2.54
    ¯
    -2.49
    ĥ
    -2.47
    ĩ
    -2.39
    ĵ
    -2.39
    Į
    -2.38
    Ŀ
    -2.32
    ł
    -2.23
    POSITIVE LOGITS
     pictures
    3.11
     photographed
    2.89
     footage
    2.87
     pictured
    2.86
     photograph
    2.85
     pics
    2.73
     painted
    2.64
     photographs
    2.63
     photo
    2.62
     photos
    2.55
    Activations Density 3.650%

    No Known Activations