© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 21155
    Prev
    Next
    INDEX
    Explanations

    want to kill or die

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

    The neuron fires on first‐person expressions of strong desires or urges (e.g. “want to…,” “kill…,” “eat…,” “jump…,” “die…”).

    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
    谞
    -1.41
    lleg
    -1.23
     REAL
    -1.23
    蘑菇
    -1.20
     grze
    -1.19
    égio
    -1.18
                           
    -1.18
     FINAL
    -1.17
    ଢ
    -1.16
    缭
    -1.16
    POSITIVE LOGITS
     only
    1.48
     all
    1.47
     just
    1.34
     everything
    1.31
     congr
    1.17
     Plung
    1.16
     both
    1.16
     always
    1.14
     スポ
    1.14
     pouvez
    1.13
    Activations Density 0.046%

    No Known Activations