© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-27B-IT
    3. 16-GEMMASCOPE-2-TRANSCODER-262K
    4. 99079
    Prev
    Next
    INDEX
    Explanations

    **Analyzing Neuron Behavior**I'm going to analyze the provided lists (MAX_ACTIVATING_TOKENS, TOKENS_AFTER_MAX_ACTIVATING_TOKEN, TOP_POSITIVE_LOGITS, TOP_ACTIVATING_TEXTS) to identify a pattern or concept represented by the neuron. My output will be a concise phrase (3 to 20 words) that describes this behavior, avoiding lists of tokens and prefacing phrases.**Analysis:*** **MAX_ACTIVATING_TOKENS**: Contains `ophe`, `ophe`, `void`, `void`, `void`, ...* **TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: Corresponds to `ophe` -> `les`, `void` -> `Start`.* **TOP_ACTIVATING_TEXTS**: * "infected *Anopheles* mosquitoes.`void Start()` or `ophe les`

    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_16_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
     prepared
    0.57
    nosti
    0.55
     P
    0.53
     aus
    0.51
    firm
    0.51
    asis
    0.50
    prepared
    0.49
     Hague
    0.48
    it
    0.48
     presumption
    0.47
    POSITIVE LOGITS
    лися
    0.52
    鳧
    0.50
    хло
    0.48
     Trên
    0.47
    ξι
    0.46
     oiseaux
    0.45
    த்த
    0.45
    രോ
    0.45
     Operator
    0.44
     সংক্রম
    0.44
    Activations Density 0.001%

    No Known Activations