© 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. 66645
    Prev
    Next
    INDEX
    Explanations

    sabotage, stockpile, boycott, troubleshoot, bomb, vent, spam

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

    the neuron detects words naming hostile or obstructive actions (e.g. sabotage, boycott, ambushed, bombing).

    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
     think
    -1.55
    </h3>
    -1.52
     メニュー
    -1.52
    re
    -1.52
     była
    -1.48
     затем
    -1.44
     Therefore
    -1.38
    會有
    -1.38
     Frequently
    -1.36
    用戶
    -1.36
    POSITIVE LOGITS
    ”？
    1.72
    lgari
    1.63
     ?
    1.62
     retten
    1.59
     kritis
    1.58
    </b>
    1.58
    沨
    1.58
     habis
    1.54
    abaikan
    1.52
    INGTON
    1.51
    Activations Density 0.038%

    No Known Activations