© 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. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 127621
    Prev
    Next
    INDEX
    Explanations

    conclusion

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron detects section‐heading or summary‐type labels (e.g. “Wrapping up,” “Conclusion,” “Epilogue,” “Roundup,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     is
    -1.53
     it
    -1.42
    ↵↵
    -1.37
    eapon
    -1.35
     viewed
    -1.30
     positioned
    -1.27
    躥
    -1.27
    èbre
    -1.27
     berguna
    -1.27
    ⦑
    -1.26
    POSITIVE LOGITS
    3
    1.56
    ir
    1.35
    </strong>
    1.34
    0
    1.30
     new
    1.30
     که
    1.27
    Sebelumnya
    1.27
    //
    1.27
    la
    1.27
    a
    1.26
    Activations Density 0.010%

    No Known Activations