© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 84333
    Prev
    Next
    INDEX
    Explanations

    should be seen or understood

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

    The neuron specifically fires on the modal verb “should,” flagging suggestion or obligation statements.

    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
     planeta
    -1.23
     during
    -1.22
    butuhkan
    -1.16
     six
    -1.15
     several
    -1.12
     zj
    -1.09
     four
    -1.09
     any
    -1.09
     will
    -1.05
     seven
    -1.05
    POSITIVE LOGITS
     tohto
    1.41
     consider
    1.27
     spôso
    1.25
    も含
    1.23
    大丈夫です
    1.20
    ないです
    1.19
     incluindo
    1.17
    ఝ
    1.17
    Consider
    1.15
    nith
    1.15
    Activations Density 0.035%

    No Known Activations