© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 6948
    Prev
    Next
    INDEX
    Explanations

    action at a distance

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

    This neuron fires on tokens that begin a sentence or block—i.e. words (often single or capitalized function words) in the sentence-initial position.

    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
    0.71
    :
    0.70
     Format
    0.66
     keys
    0.59
     Informed
    0.56
    #
    0.55
     communiqué
    0.54
    ],
    0.54
    }:
    0.53
    },
    0.52
    POSITIVE LOGITS
    aunque
    0.82
     capacit
    0.80
     menores
    0.77
    However
    0.75
    doesn
    0.75
    Although
    0.74
    isn
    0.74
    Ꮘ
    0.73
     reproduct
    0.73
     puedan
    0.72
    Activations Density 0.000%

    No Known Activations