© 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. 50162
    Prev
    Next
    INDEX
    Explanations

    it is or it can

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

    The neuron strongly responds to common function words acting as referential or transitional markers (e.g. “it,” “this,” “which,” “is,” “in addition,” “could,” “you”).

    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
    に行った
    -0.98
     стаття
    -0.98
    場合には
    -0.98
     选择
    -0.95
    過程中
    -0.93
     chooses
    -0.93
    אני
    -0.92
    なのだ
    -0.92
     chose
    -0.91
    绶
    -0.91
    POSITIVE LOGITS
     can
    2.09
     will
    2.02
     is
    1.71
     also
    1.56
     could
    1.46
     are
    1.16
    reportWebVitals
    1.10
     would
    1.09
    สามารถ
    1.05
    CIUDAD
    1.01
    Activations Density 0.004%

    No Known Activations