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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B-IT
    5. 40-GEMMASCOPE-2-RES-262K
    7. 12221
    Prev
    Next
    INDEX
    Explanations

    "that's a question"

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

    Whether a token is part of the model/assistant's generated response (i.e., detects model/assistant speech turns).

    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_40_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
     propone
    0.47
     tycker
    0.41
     acudir
    0.40
     мнению
    0.40
    }|$
    0.40
    👎
    0.39
    ವಿದೆ
    0.39
    削
    0.38
    垫
    0.38
    醮
    0.38
    POSITIVE LOGITS
     complicated
    0.62
     peculiar
    0.58
     confusing
    0.56
    strange
    0.55
     strange
    0.54
    complicated
    0.54
     technically
    0.53
     confuses
    0.52
     awkward
    0.51
     weird
    0.49
    Activations Density 0.030%

    No Known Activations