© 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. 31-GEMMASCOPE-2-RES-16K
    7. 70
    Prev
    Next
    INDEX
    Explanations

    introduces clauses starting with "that"

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

    This neuron responds to assertive reporting or stance-taking verbs (e.g. ensure, say/said, argue, clear, make, estimates).

    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
    +
    0.36
    그리고
    0.35
    あるいは
    0.35
    そして
    0.33
    Layout
    0.32
    などを
    0.32
    =
    0.31
    ／
    0.30
    ？）
    0.30
    以及
    0.30
    POSITIVE LOGITS
     that
    1.63
    that
    1.49
     bahwa
    1.22
     rằng
    1.22
     THAT
    1.13
     bahawa
    1.12
     दैट
    1.08
     thats
    1.07
     दट
    1.05
    THAT
    1.05
    Activations Density 1.090%

    No Known Activations