© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 2575
    Prev
    Next
    INDEX
    Explanations

    must followed by verbs of obligation or understanding

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

    The main thing this neuron does is find words and phrases that express obligation or necessity (e.g., must, need to, have to).

    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
     köpa
    -1.09
     will
    -1.08
     []*
    -1.05
    aneity
    -1.03
     referrerpolicy
    -1.02
    :'',
    -1.00
    ₒ
    -0.95
    achieve
    -0.93
     atrav
    -0.92
    nextElement
    -0.92
    POSITIVE LOGITS
     remember
    2.81
    remember
    2.22
     understand
    2.20
    Remember
    2.11
     remembers
    2.09
     remembered
    1.89
     pamię
    1.73
     Remember
    1.71
     remembering
    1.63
     ensure
    1.61
    Activations Density 0.030%

    No Known Activations