© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 130602
    Prev
    Next
    INDEX
    Explanations

    ResponseStatus, toList, anthologized, Muckraker, anonymized

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

    The neuron fires on the very first subword piece of longer words (i.e. the initial BPE fragment at the start of multi‐syllable, camelCase, or compound tokens).

    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
    .”
    -3.02
    %;
    -2.89
    We
    -2.73
    )
    -2.55
    ра
    -2.52
    ༵
    -2.48
    L
    -2.45
    .
    -2.34
    ")]
    -2.33
    𓋼
    -2.33
    POSITIVE LOGITS
    ers
    2.66
    código
    2.53
     withered
    2.53
    FOTOS
    2.45
     dichas
    2.45
     multitude
    2.38
     deterred
    2.36
     murieron
    2.31
     forefront
    2.28
     тогда
    2.27
    Activations Density 0.005%

    No Known Activations