© 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. 67252
    Prev
    Next
    INDEX
    Explanations

    design intention

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

    The neuron consistently activates on mentions of “intention” (and related forms like “intend”), i.e. words indicating someone’s intent.

    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
    寕
    -2.48
     envelop
    -2.45
     increíbles
    -2.42
    pulseira
    -2.25
     differing
    -2.25
    ޏ
    -2.22
    تیجه
    -2.20
    廕
    -2.19
    ㈲
    -2.17
    M
    -2.14
    POSITIVE LOGITS
     That
    2.52
    ),
    2.47
    1
    2.38
    “
    2.27
    鞴
    2.14
    な感じ
    2.08
     You
    2.05
    you
    2.00
    ist
    1.98
     Those
    1.95
    Activations Density 0.010%

    No Known Activations