© 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. 31-GEMMASCOPE-2-RES-16K
    4. 289
    Prev
    Next
    INDEX
    Explanations

    The neuron is primarily detecting personal‐address pronouns and contractions—especially first‐ and second‐person references like “I,” “you,” “’ve,” “we,” etc.

    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
    😍
    0.46
    つなが
    0.46
    During
    0.45
    庆祝
    0.45
     themed
    0.45
    ☀️
    0.44
    带动
    0.44
    ✨
    0.43
    たくさん
    0.43
    能够
    0.43
    POSITIVE LOGITS
     illogical
    1.48
     ineffective
    1.46
     inadequate
    1.44
     unrealistic
    1.44
     inappropriate
    1.42
     pointless
    1.42
     untenable
    1.42
     inaccurate
    1.41
     unjustified
    1.41
     meaningless
    1.41
    Activations Density 5.828%

    No Known Activations