© 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
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 8978
    Prev
    Next
    INDEX
    Explanations

    emojis andgiphy

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron detects emoticon-style tokens (e.g. Slack/IRC codes like :slight_smile: and asterisk‐enclosed actions such as *high fives* ).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    emoticons and text-based expressions of emotion or reactions (like ":eek:", "*high fives*", ":(", "*mad giggle*").

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    emoji/emoticon-style expressions and action-aside gestures indicated by punctuation delimiters (such as colon- or asterisk-bounded phrases and ASCII emoticons).

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Copyright
    0.67
     N
    0.62
     Blockchain
    0.60
     None
    0.57
     none
    0.57
     |
    0.56
     Nature
    0.54
     J
    0.53
     News
    0.53
     Nest
    0.53
    POSITIVE LOGITS
     emoji
    1.07
     emojis
    1.07
    emoji
    1.05
     emoticon
    1.05
     shrug
    0.98
     sobbing
    0.94
    Emoji
    0.91
    giphy
    0.89
     gestures
    0.86
     Emoji
    0.86
    Activations Density 0.047%

    No Known Activations