© 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. 34-GEMMASCOPE-RES-131K
    4. 65768
    Prev
    Next
    INDEX
    Explanations

    can/to followed by action

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

    The neuron detects user requests or questions by activating on words and tokens signaling asking for help (e.g. “please,” “how,” “can,” “anyone,” “want to,” etc.).

    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
    𝟑
    -1.00
     suplementos
    -0.99
     capítulos
    -0.99
    TWO
    -0.97
     queremos
    -0.96
     heeft
    -0.94
    lusst
    -0.94
     estimés
    -0.93
    في
    -0.93
    ldquo
    -0.93
    POSITIVE LOGITS
     aware
    1.12
     anyhow
    1.11
     involve
    1.07
    osoever
    0.93
     bifur
    0.92
     fulfill
    0.90
     flickr
    0.90
     give
    0.90
     周辺
    0.88
     adjoint
    0.87
    Activations Density 0.184%

    No Known Activations