© 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. 10876
    Prev
    Next
    INDEX
    Explanations

    question to model response

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

    The neuron fires strongly on the first content word of a new topic or paragraph—i.e. it detects beginnings of major text blocks.

    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
    ז
    0.98
    զ
    0.96
    ч
    0.94
    з
    0.89
    ר
    0.88
    אַר
    0.87
    ἥ
    0.87
    нд
    0.84
     SLAs
    0.81
    lini
    0.80
    POSITIVE LOGITS
     conjunta
    0.78
    argue
    0.73
     agrade
    0.68
     innovación
    0.67
    }];
    0.66
     asimismo
    0.66
    órios
    0.66
     automática
    0.66
    profitable
    0.66
     съм
    0.65
    Activations Density 0.000%

    No Known Activations