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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 11050
    Prev
    Next
    INDEX
    Explanations

    actions happening

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

    The main thing this neuron does is detect references to hostile actors or perpetrators of violence and crime.

    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
     which
    -0.98
    конча
    -0.98
     förb
    -0.98
     vueltas
    -0.95
     pemas
    -0.93
     melainkan
    -0.92
     ihop
    -0.92
     how
    -0.92
    および
    -0.91
    それが
    -0.91
    POSITIVE LOGITS
     themselves
    1.43
    Bigger
    1.22
    に行きました
    1.20
    appliquent
    1.13
    бовал
    1.12
     their
    1.09
     ingress
    1.09
     Técnicas
    1.09
    medik
    1.08
    Worse
    1.07
    Activations Density 0.048%

    No Known Activations