© 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. 617
    Prev
    Next
    INDEX
    Explanations

    questions and exercises

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

    The neuron primarily fires on numeric tokens and other metadata markers (e.g. the “Q” line and up‐vote or ID numbers) that appear at the top of a StackOverflow–style question.

    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
     Ciò
    -0.80
     malang
    -0.78
     Comuni
    -0.76
    decap
    -0.76
    ędziesz
    -0.76
     bendera
    -0.75
    quín
    -0.75
     répartition
    -0.74
    ޅ
    -0.74
    houettes
    -0.73
    POSITIVE LOGITS
     question
    3.88
     Question
    3.66
    Question
    3.05
     questions
    2.86
    question
    2.84
    QUESTION
    2.53
     QUESTION
    2.48
     Exercise
    2.39
     part
    2.36
     exercise
    2.36
    Activations Density 0.247%

    No Known Activations