© 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. 15238
    Prev
    Next
    INDEX
    Explanations

    instances of the word "try" in various contexts

    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
    489
    +0.14
    0.8%
    377
    +0.14
    0.8%
    279
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    489
    +0.14
    0.04
    377
    +0.14
    0.04
    286
    +0.12
    0.04
    Negative Logits
    Ļª
    -2.37
    ĥ½
    -2.35
    ĨĴ
    -2.28
    ĺ
    -2.10
    ·¸
    -2.00
    ļ
    -1.97
    ŀ
    -1.96
    ľĵ
    -1.94
    ĥ
    -1.92
    Ĩ
    -1.89
    POSITIVE LOGITS
    psin
    2.35
     unsuccess
    2.22
     desperately
    2.19
     again
    2.12
     harder
    2.12
    outs
    1.94
     to
    1.88
     contacting
    1.79
    ste
    1.75
    out
    1.74
    Activations Density 0.200%

    No Known Activations