© 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
    3. Gemma-3-12B-IT
    5. 12-GEMMASCOPE-2-RES-16K
    7. 14501
    Prev
    Next
    INDEX
    Explanations

    **Identifying Commas and Following Words**I've analyzed the provided lists. The neuron is strongly activated by commas. The `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` list shows a variety of words that appear immediately after these commas. The `TOP_POSITIVE_LOGITS` list presents tokens from various languages, suggesting the neuron's activations might not be tied to a single language but rather to the structural element itself—the comma.Therefore, the clearest pattern is the recognition of commas, followed by diverse content.**Explanation:**commas and following text

    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_12_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
    lja
    1.40
     conceivably
    1.26
    tipo
    1.25
    ssymb
    1.24
    vih
    1.20
    natures
    1.20
    gallery
    1.18
    cemic
    1.18
    camore
    1.17
    ndor
    1.17
    POSITIVE LOGITS
     również
    1.45
    い
    1.41
    其他
    1.23
    और
    1.22
     i
    1.18
    IA
    1.18
    การ
    1.16
     Downloads
    1.16
    od
    1.15
    het
    1.14
    Activations Density 0.040%

    No Known Activations