© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5176
    Prev
    Next
    INDEX
    Explanations

    preventing harm or damage

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

    This neuron detects mentions of damage, injury, harm, or other adverse‐event terms (including liability/compensation language).

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    ('#
    0.71
     沈
    0.70
    ("#
    0.68
    FragmentManager
    0.63
    呛
    0.62
    LongNumber
    0.60
    بہ
    0.60
    LOTREntity
    0.59
     ansch
    0.59
     Entire
    0.59
    POSITIVE LOGITS
     damage
    1.93
     Damage
    1.82
     DAMAGE
    1.70
    Damage
    1.64
    damage
    1.62
     harm
    1.58
     daño
    1.48
    损伤
    1.48
     damages
    1.46
    伤害
    1.43
    Activations Density 0.209%

    No Known Activations