© 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. 2318
    Prev
    Next
    INDEX
    Explanations

    fire, firewalls, firefly

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

    The neuron strongly activates on the word “fire” and other “fire”-prefixed terms (e.g. firewall, FireEye, Fire Nation).

    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
     envers
    0.67
    dicts
    0.58
    光滑
    0.58
    毉
    0.57
    憎
    0.56
    绞
    0.55
     debtors
    0.54
     Doo
    0.54
     Tension
    0.53
     jīn
    0.53
    POSITIVE LOGITS
    Fire
    1.43
     Fire
    1.36
     fire
    1.35
     firefighting
    1.34
     extinguisher
    1.33
    FIRE
    1.31
    fire
    1.30
    🔥🔥
    1.28
     FIRE
    1.27
     फायर
    1.26
    Activations Density 0.168%

    No Known Activations