© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 22410
    Prev
    Next
    INDEX
    Explanations

    configurable, OutOfRange, and Korean usage

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

    The neuron flags uncommon or “complex” tokens—long/rare English terms (e.g. configurable, OutOfRange) as well as non-English script tokens (Japanese, Chinese, Korean).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    かなかった
    -2.45
                    
    -2.41
    人たち
    -2.27
    </b>
    -2.22
     двух
    -2.17
    なった
    -2.13
     dejando
    -2.09
    甦
    -2.09
     hacemos
    -2.00
    好似
    -2.00
    POSITIVE LOGITS
    酲
    2.83
    ൎ
    2.64
    洐
    2.59
     beslu
    2.50
    ꪔ
    2.47
    ternut
    2.38
    熥
    2.31
    扃
    2.31
    媜
    2.31
    ⸒
    2.25
    Activations Density 0.001%

    No Known Activations