© 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-3-27B-IT
    5. 31-GEMMASCOPE-2-RES-262K
    7. 2017
    Prev
    Next
    INDEX
    Explanations

    would you or be

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

    sentences or phrases where the speaker (the assistant) offers help, expresses willingness, or asks what the user would like.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_31_width_262k_l0_medium
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     (
    0.60
    ika
    0.55
     When
    0.52
     hints
    0.50
     I
    0.49
     are
    0.49
     ต
    0.46
     є
    0.46
     Because
    0.46
     by
    0.46
    POSITIVE LOGITS
     avrebbe
    0.77
     tendría
    0.77
     haría
    0.75
     estaría
    0.70
    erebbe
    0.69
     podría
    0.66
     conceivably
    0.66
     sarebbe
    0.66
     habría
    0.65
     sería
    0.64
    Activations Density 0.182%

    No Known Activations