© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B-IT
    5. 31-GEMMASCOPE-2-RES-262K
    7. 81
    Prev
    Next
    INDEX
    Explanations

    model response generation

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

    the assistant/model speaker tag in chat transcripts, especially when followed by a numeric identifier.

    oai_token-act-pair · gpt-5Triggered 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
     raies
    0.30
     وجہ
    0.30
    mensaje
    0.29
     filles
    0.29
    rlig
    0.28
    CCESS
    0.28
     ಏಕೆ
    0.28
     halibut
    0.28
    峴
    0.28
    netHex
    0.28
    POSITIVE LOGITS
    Detailed
    0.28
     Detailed
    0.28
     LISTEN
    0.27
    Ideally
    0.27
     चाहते
    0.26
     Elabor
    0.25
     ডেভেলপ
    0.25
    二次
    0.25
    全面的
    0.25
     détaillée
    0.25
    Activations Density 0.173%

    No Known Activations