© 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. 15-GEMMASCOPE-2-TRANSCODER-262K
    4. 17810
    Prev
    Next
    INDEX
    Explanations

    The `Turn` token appears twice in `MAX_ACTIVATING_TOKENS`, and `itin` (suggesting "Turnitin") appears twice in `TOKENS_AFTER_MAX_ACTIVATING_TOKEN`.The `similarity` token also appears twice in `MAX_ACTIVATING_TOKENS`.`TOP_ACTIVATING_TEXTS` contains mentions of "Turnitin", "plagiarism checking software", "similarity", "AI detection tools", "perplexity", "copied code", "exact matches", and "MOSS".The common theme is identifying or checking for similarity, plagiarism, or copied content, often using tools like Turnitin.Let's consider the patterns:- "Turn" followed by "itin" (implied) relates to plagiarism checking tools.- "similarity" is directly related to plagiarism and content matching.- The other tokens in `TOP_POSITIVE_LOGITS` are diverse and mostly unrelated common words or technical terms in different languages, which might be noise or indicate a broad reach but not a specific common pattern for this neuron's *core* function.**plagiarism and similarity detection**

    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_15_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
    1
    0.71
    ↵↵
    0.63
    3
    0.56
                
    0.55
    …
    0.53
     in
    0.52
     {
    0.52
    /
    0.52
     
    0.51
     dalam
    0.50
    POSITIVE LOGITS
     Стаўкі
    0.56
     récord
    0.55
    𝑢
    0.55
    сек
    0.54
     ფედერ
    0.52
    PlistBuddy
    0.52
    टर्न
    0.52
    ਟਰ
    0.52
    病人
    0.52
    мата
    0.51
    Activations Density 0.000%

    No Known Activations