© 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. 128942
    Prev
    Next
    INDEX
    Explanations

    The neuron is detecting words that describe hostile or violent actions (e.g. attack, invaders, raids, assassins).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    attacks, predators, invasion

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

    references to conflict, attacks, and threats in text.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi
    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
     ну
    -0.96
    lasyon
    -0.92
     Той
    -0.90
    ︰
    -0.90
    יח
    -0.87
     eget
    -0.86
    stappen
    -0.86
     behov
    -0.86
    StreetMap
    -0.85
    ultas
    -0.84
    POSITIVE LOGITS
     threat
    1.20
     they
    1.17
    threat
    1.03
     threats
    0.98
     amenazas
    0.97
     protection
    0.95
     attack
    0.95
     mentre
    0.94
     protezione
    0.91
     ellos
    0.90
    Activations Density 0.078%

    No Known Activations