© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 12877
    Prev
    Next
    INDEX
    Explanations

    Japanese characters and English prepositions

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

    The neuron strongly activates on isolated single-letter tokens (especially uppercase letters), i.e. standalone initials.

    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_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
    భర
    0.13
    ሁሉም
    0.12
    ፔ
    0.11
    এদিকে
    0.11
     đủ
    0.11
    सभी
    0.11
     স্ত্রীকে
    0.11
    Hepin
    0.11
    캄
    0.10
    itories
    0.10
    POSITIVE LOGITS
     F
    0.13
     S
    0.12
     Watercolor
    0.12
     J
    0.12
     Refining
    0.12
     Science
    0.12
     Sci
    0.12
     The
    0.11
     M
    0.11
     Water
    0.11
    Activations Density 0.000%

    No Known Activations