© 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
    2. Gemma-3-27B-IT
    3. 10-GEMMASCOPE-2-TRANSCODER-262K
    4. 48389
    Prev
    Next
    INDEX
    Explanations

    The user wants a concise explanation (3-20 words) of a neuron's behavior, based on patterns found in four lists.Let's analyze the lists:1. **MAX_ACTIVATING_TOKENS**: `1, 1, 2, 1, ik, 2, 2, dist, 1, the` * Contains numbers (1, 2) and common words ('the'). Also 'ik' and 'dist'.2. **TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: `9, -, game, -, af, *. , 4, orts, 0, preferred` * Words like 'game' and 'preferred' appear.3. **TOP_POSITIVE_LOGITS**: `ne, which, ki, id, that, in, which, in, from, ma` * Common words and potential word beginnings/endings.4. **TOP_ACTIVATING_TEXTS**: * Mix of code snippets, technical explanations, and conversational text. * One text mentions "CS1955", "CS1-DQ", "CS1-I2C", "CS2 game process", "CS2. Rank S", "CS240 server", "CS109", "CS50’s".cs1, cs2, cs109, cs240

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_10_width_262k_l0_small_affine
    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
    地方
    0.60
     nostre
    0.57
    お得
    0.56
    朋友
    0.54
    я
    0.53
    不
    0.53
     nossos
    0.53
     vijf
    0.52
     échange
    0.52
    వ
    0.52
    POSITIVE LOGITS
    ne
    0.80
     which
    0.77
    ki
    0.76
    id
    0.74
     that
    0.72
     in
    0.71
    which
    0.71
    in
    0.71
     from
    0.71
    ma
    0.71
    Activations Density 0.000%

    No Known Activations