© 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. 98476
    Prev
    Next
    INDEX
    Explanations

    accelerating tube

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

    The neuron selectively fires on subword fragments of “accelerate”/“accelerator” (e.g. the “ac-” and “celer” pieces of those words).

    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
    ÁCT
    -1.24
    Transk
    -1.23
    柘
    -1.21
    古屋
    -1.20
    石垣
    -1.19
    筧
    -1.16
    Она
    -1.13
     instala
    -1.13
    ŚĆ
    -1.13
    geri
    -1.12
    POSITIVE LOGITS
    <b>
    1.29
    ल
    1.06
     alojamientos
    1.05
    mathbf
    1.03
    我
    1.02
    Chapter
    1.02
    V
    1.01
    三
    1.01
    ह
    1.01
    X
    0.99
    Activations Density 0.011%

    No Known Activations