© 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. 8535
    Prev
    Next
    INDEX
    Explanations

    references to set notions or elements within a context

    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
    194
    +0.13
    0.7%
    78
    +0.11
    0.6%
    263
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    194
    +0.13
    0.04
    462
    +0.11
    0.03
    160
    +0.11
    0.02
    Negative Logits
    ingo
    -1.74
    UTERS
    -1.74
     DISTRICT
    -1.71
    ILY
    -1.54
    =>
    -1.42
    icy
    -1.39
    pled
    -1.37
    $’
    -1.36
    woke
    -1.31
     sake
    -1.30
    POSITIVE LOGITS
     aside
    3.02
    ups
    2.98
     forth
    2.83
     up
    2.68
     Aside
    2.27
    ters
    2.16
    backs
    2.13
    tee
    2.12
     apart
    2.10
    upt
    2.07
    Activations Density 0.132%

    No Known Activations