© 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. 130714
    Prev
    Next
    INDEX
    Explanations

    The neuron detects language expressing recommendations or advice (e.g. “recommend,” “strongly recommended,” “highly recommend”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    recommend, appreciate, or express importance

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     enough
    -1.41
     maybe
    -1.34
    struzioni
    -1.33
     måske
    -1.33
     ocurrido
    -1.33
    轶
    -1.32
     perhaps
    -1.30
     adecuados
    -1.29
     adecuado
    -1.29
     mainly
    -1.29
    POSITIVE LOGITS
    鿬
    1.17
     própri
    1.16
     it
    1.13
     that
    1.13
    ఝ
    1.12
    toit
    1.11
     userEmail
    1.08
    が
    1.08
     professionelle
    1.07
     medan
    1.06
    Activations Density 0.048%

    No Known Activations