© 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. 31-GEMMASCOPE-2-RES-16K
    4. 10487
    Prev
    Next
    INDEX
    Explanations

    data.info() or data.describe()

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

    The neuron spikes on discourse-linking cue words that introduce examples, contrasts, or sources—i.e. transition markers like “such,” “Not only,” “According to,” etc.

    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
     erros
    0.70
     errors
    0.66
     correspondance
    0.62
     JN
    0.59
    뻑
    0.59
     mistakes
    0.58
     erreurs
    0.58
     pegs
    0.57
     uplifting
    0.56
     Libr
    0.56
    POSITIVE LOGITS
    ))^{
    0.70
    })^{
    0.69
    Batch
    0.68
    })^
    0.68
    ")).
    0.61
    ))^
    0.60
    Bounded
    0.60
    Sql
    0.60
    }|^
    0.59
    Gen
    0.59
    Activations Density 0.006%

    No Known Activations