© 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-3-12B-IT
    3. 24-GEMMASCOPE-2-RES-16K
    4. 1836
    Prev
    Next
    INDEX
    Explanations

    The neuron seems to activate on nouns that often appear at the end of a sentence or list item. The `TOP_POSITIVE_LOGITS` also heavily feature punctuation, especially closing quotation marks and symbols like ® and ™.Let's look for a unifying theme. The `MAX_ACTIVATING_TOKENS` are all nouns that often represent distinct entities, categories, or concepts: collectibles, products, groups, pages, perspectives, treasure, industries, creatures, challenges. The fact that they are often followed by a period or comma suggests they are frequently the terminal element of a phrase or sentence.The `TOP_POSITIVE_LOGITS` being punctuation strongly supports this. The presence of ® and ™ symbols could imply that some of these nouns are brand names or trademarked entities.Considering the `TOP_ACTIVATING_TEXTS`:* "Living Assets," dynamic and scalable NFTs that evolve based on usage, offering enhanced utility and ownership beyond static collectibles.**The neuron detects nouns often ending phrases or lists, sometimes with trademark symbols.**

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-it/resid_post/layer_24_width_16k_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
     orange
    0.67
     strawberry
    0.64
     거고
    0.63
     если
    0.60
     if
    0.59
     snappy
    0.59
     cheesy
    0.57
     towel
    0.57
     an
    0.56
     cheese
    0.56
    POSITIVE LOGITS
    ."
    0.85
    ®.
    0.85
    .”
    0.84
    ™.
    0.83
     poprzez
    0.77
    。”
    0.77
    .’
    0.76
    .\"
    0.76
     attraverso
    0.73
    .“
    0.72
    Activations Density 0.000%

    No Known Activations