© 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. 24-GEMMASCOPE-2-RES-16K
    4. 5563
    Prev
    Next
    INDEX
    Explanations

    question prompts

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

    This neuron strongly activates on question‐format tokens—especially blank placeholders (“__”) and numeric question labels/numbers.

    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
     当然
    0.59
    }}\|
    0.55
    }):
    0.55
    也将
    0.55
    íř
    0.54
    然后再
    0.53
     също
    0.52
     дальнейшем
    0.51
    睉
    0.50
     böyle
    0.50
    POSITIVE LOGITS
     Which
    0.84
     _______
    0.84
     _____
    0.82
     ___
    0.80
     ______
    0.80
     __________
    0.80
     ____
    0.79
    Which
    0.77
     ________
    0.75
    what
    0.74
    Activations Density 0.013%

    No Known Activations