© 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. 32516
    Prev
    Next
    INDEX
    Explanations

    email addresses

    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
    74
    +0.13
    0.7%
    185
    +0.13
    0.7%
    317
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    74
    +0.13
    0.02
    185
    +0.13
    0.02
    298
    +0.13
    0.02
    Negative Logits
     necessary
    -1.77
     thereto
    -1.69
     due
    -1.65
     satisfying
    -1.63
     sampled
    -1.62
     outper
    -1.62
    .’”
    -1.61
     maxim
    -1.58
    lessly
    -1.58
     exactly
    -1.57
    POSITIVE LOGITS
    ®
    2.12
    ª
    2.10
    ±
    2.05
    ¦
    2.04
    ¾
    2.03
    cdn
    2.01
    «
    1.99
    ¡
    1.94
    ¬
    1.92
    Ń
    1.91
    Activations Density 0.151%

    No Known Activations