© 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. 65525
    Prev
    Next
    INDEX
    Explanations

    targeting victims or weaknesses

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

    The neuron is looking for nouns or noun phrases that denote potential targets or objects of attack, stealth, or acquisition in hostile contexts.

    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
    IBRATION
    -0.97
     we
    -0.96
    手里
    -0.93
    用到
    -0.91
    劄
    -0.86
     situa
    -0.86
     obliga
    -0.86
    wiście
    -0.85
     resina
    -0.84
    الحياه
    -0.84
    POSITIVE LOGITS
     vulnerable
    1.16
     vulnerabilities
    1.12
    Targeting
    1.04
     targets
    1.02
     weaknesses
    1.02
     targeted
    0.99
     target
    0.99
     targeting
    0.96
     victim
    0.94
     weakness
    0.94
    Activations Density 0.111%

    No Known Activations