© 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. 39294
    Prev
    Next
    INDEX
    Explanations

    asking for any help

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

    The neuron fires strongly on the common “Thanks for any help” closing (especially the words “for,” “any,” and “help”), i.e. it spotlights requests for assistance.

    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
     a
    -0.98
    েম
    -0.93
    зайн
    -0.91
     мно
    -0.90
    gameOver
    -0.89
     kek
    -0.85
     the
    -0.85
    ETCH
    -0.85
    retty
    -0.84
     perlu
    -0.84
    POSITIVE LOGITS
     this
    1.34
     any
    1.24
     will
    1.23
     each
    1.20
     help
    1.16
    每個
    1.11
     каждого
    1.09
    Each
    1.08
     cualquier
    1.00
    estias
    1.00
    Activations Density 0.019%

    No Known Activations