© 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-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 33318
    Prev
    Next
    INDEX
    Explanations

    handle or avoid

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

    The neuron flags numeric tokens—i.e. standalone numbers or measurements—in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    んですが
    -1.23
    ⠀⠀⠀⠀
    -1.20
    ⠀⠀
    -1.18
     peinados
    -1.18
    んですけど
    -1.17
    antaranya
    -1.15
     faisant
    -1.12
    のですが
    -1.11
    љ
    -1.09
    wouldn
    -1.08
    POSITIVE LOGITS
     avoid
    2.67
     ensure
    2.55
     never
    2.50
     use
    2.38
     if
    2.31
     always
    2.22
     don
    1.98
     make
    1.96
     do
    1.93
     keep
    1.93
    Activations Density 0.026%

    No Known Activations