© 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. 32354
    Prev
    Next
    INDEX
    Explanations

    elements related to a storyline or narrative development

    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
    47
    +0.18
    1.0%
    56
    +0.16
    0.9%
    186
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    56
    +0.18
    -0.06
    496
    +0.16
    0.14
    392
    +0.15
    0.16
    Negative Logits
    ľ
    -1.69
     coverage
    -1.67
    \)
    -1.65
     publication
    -1.58
    ķ
    -1.53
     NEWS
    -1.51
    ↵Č
    -1.50
    quartile
    -1.49
     Bernstein
    -1.48
     Friedman
    -1.47
    POSITIVE LOGITS
     sword
    2.20
     dunge
    2.19
     Dragons
    1.93
     armor
    1.83
     enemy
    1.82
     stones
    1.81
     monsters
    1.81
     demons
    1.78
     swords
    1.78
     Evil
    1.78
    Activations Density 7.470%

    No Known Activations