© 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
    3. 24-GEMMASCOPE-2-RES-16K
    4. 375
    Prev
    Next
    INDEX
    Explanations

    The neuron flags adjectives describing awkwardness or difficulty (e.g. “awkward,” “challenging,” “difficult,” “frustrating,” “strange,” “disconcerting”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_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
     excellent
    0.50
     excell
    0.49
    excellent
    0.47
     selles
    0.44
     средней
    0.43
    👍
    0.43
     Excellent
    0.43
     nhiệt
    0.42
    ermanfaat
    0.42
     grazie
    0.42
    POSITIVE LOGITS
    StringTo
    0.50
    不已
    0.49
    sometimes
    0.49
    悚
    0.48
    Sometimes
    0.47
    Emp
    0.47
    Mouse
    0.47
    你
    0.46
     blatantly
    0.46
     अभिकथन
    0.46
    Activations Density 0.000%

    No Known Activations