© 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. 34-GEMMASCOPE-RES-131K
    7. 115601
    Prev
    Next
    INDEX
    Explanations

    harm

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

    The neuron fires on mentions of “harm” (and its close variants like “harmful”) indicating negative impact or damage.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     salvador
    -1.02
     lavabo
    -0.98
     bleus
    -0.96
    ţie
    -0.93
     paloma
    -0.92
    儺
    -0.92
    ábor
    -0.91
    ziału
    -0.91
    emma
    -0.89
    ンの
    -0.86
    POSITIVE LOGITS
     harm
    1.79
    fully
    1.20
     done
    1.20
    Harm
    1.14
    ful
    1.13
     damage
    1.10
     здоров
    1.09
     daño
    1.05
    fulness
    1.04
     to
    1.03
    Activations Density 0.013%

    No Known Activations