© 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. 12289
    Prev
    Next
    INDEX
    Explanations

    as discussed/described/mentioned above

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

    The neuron detects words and short phrases that point back to earlier parts of the text (e.g. “listed,” “above,” “previous paragraph,” “noted above”).

    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
     shown
    0.63
    shown
    0.59
     모습을
    0.58
     сказали
    0.57
     बताए
    0.56
     соль
    0.56
     بتایا
    0.54
    몫
    0.54
     сообщили
    0.53
     معلوم
    0.52
    POSITIVE LOGITS
     earlier
    0.87
     herein
    0.85
     explicitly
    0.75
     elsewhere
    0.66
     dále
    0.65
     Earlier
    0.64
     concisely
    0.64
     separately
    0.63
     above
    0.63
     trong
    0.63
    Activations Density 0.197%

    No Known Activations