© 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. 26305
    Prev
    Next
    INDEX
    Explanations

    HTML or XML tags and their attributes

    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
    28
    +0.11
    0.6%
    63
    +0.11
    0.6%
    157
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    509
    +0.11
    0.03
    28
    +0.11
    0.02
    460
    +0.11
    0.02
    Negative Logits
    cock
    -1.46
    lied
    -1.42
    yme
    -1.41
    borough
    -1.36
    imately
    -1.36
    ually
    -1.36
    bury
    -1.35
    icum
    -1.33
    plots
    -1.32
    ense
    -1.30
    POSITIVE LOGITS
    ĻĤ
    2.32
    ½
    2.24
    ¦
    2.23
    ±
    2.13
    į
    2.02
    ¿
    2.00
    ©
    1.97
    ĭ
    1.93
    IJ
    1.87
    ¥
    1.77
    Activations Density 0.129%

    No Known Activations