© 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. 22-GEMMASCOPE-RES-131K
    7. 16141
    Prev
    Next
    INDEX
    Explanations

    potential prison sentences

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

    The neuron detects mentions of criminal sentencing details—especially prison terms and their durations (e.g. years, life sentence, sentencing verbs and numbers).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     to
    -1.17
     cubana
    -1.16
     ὁ
    -1.16
     молока
    -1.06
    tiro
    -1.04
    ୀ
    -1.04
    ۞
    -1.03
     kaca
    -0.99
    她也
    -0.98
    TableBody
    -0.98
    POSITIVE LOGITS
     potentially
    1.34
    ťou
    1.23
     podrían
    1.16
     depending
    1.13
     potential
    1.13
    IDING
    1.11
     nuevo
    1.07
     possible
    1.07
    根據
    1.06
     could
    1.05
    Activations Density 0.008%

    No Known Activations