© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 68034
    Prev
    Next
    INDEX
    Explanations

    Caffeinated, exfoliation, antibiotic, fermionic

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

    chemical terms or compounds with syllables like "caffe", "ferm", "quet", and "han".

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron activates on tokens related to “caffe-” (e.g. caffeinated soda, caffeic acid), i.e. occurrences of the “caffe” root.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Ꭳ
    -3.78
    鳎
    -3.38
    ⌵
    -3.23
    饪
    -3.19
    ﮨ
    -3.14
    琑
    -3.00
    嗩
    -2.95
    ⟝
    -2.91
    銕
    -2.86
    鸲
    -2.86
    POSITIVE LOGITS
    *
    3.38
    i
    3.28
    <td>
    3.17
    "
    2.77
    ...
    2.47
    ----------------
    2.41
     it
    2.34
    ↵↵
    2.33
     slew
    2.31
    <
    2.31
    Activations Density 0.037%

    No Known Activations