© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 12385
    Prev
    Next
    INDEX
    Explanations

    that follows relative clauses

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

    The neuron strongly activates on formal legal and judicial nouns (e.g. courts, appellant, witnesses, allegations, findings).

    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
     (\"
    0.78
    —.
    0.75
     (“
    0.73
    —”
    0.69
     ().
    0.69
     ("
    0.67
     —
    0.66
     thóc
    0.66
     까
    0.64
     (£
    0.64
    POSITIVE LOGITS
     который
    1.01
     although
    1.00
     которые
    0.98
     kterou
    0.96
     которое
    0.96
     aunque
    0.95
     которая
    0.94
     welche
    0.91
     która
    0.91
    which
    0.90
    Activations Density 0.243%

    No Known Activations