© 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. 117909
    Prev
    Next
    INDEX
    Explanations

    spectral tones, high gain

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

    The neuron responds to technical RF‐engineering jargon—especially multiword terms naming modulation schemes (e.g. “multi‐level modulation,” “quadrature amplitude modulation”) and amplifier classes (e.g. “Class A,” “Class AB”).

    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
    勇士
    -0.73
    玄
    -0.72
    引っ
    -0.72
     greets
    -0.71
    loway
    -0.71
    голо
    -0.69
     rencont
    -0.69
     subgroups
    -0.68
    movieId
    -0.68
    gert
    -0.67
    POSITIVE LOGITS
    Adjacent
    0.92
     Adjacent
    0.84
    OURI
    0.84
     tones
    0.81
     adjacent
    0.80
    adjacent
    0.80
     fout
    0.79
    ASING
    0.77
     Degradation
    0.77
     constellation
    0.75
    Activations Density 0.055%

    No Known Activations