© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 1684
    Prev
    Next
    INDEX
    Explanations

    glass and subsequent word

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

    The neuron reacts strongly to occurrences of the word “glass” (in any capitalization or as part of compound names).

    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
     vad
    -0.92
    Descending
    -0.85
    決済
    -0.84
    ͘
    -0.81
    惶
    -0.80
     sacrificing
    -0.78
    Roz
    -0.77
     یاد
    -0.77
     admit
    -0.76
     कृ
    -0.76
    POSITIVE LOGITS
    work
    1.02
     shattering
    1.01
     panes
    1.01
     Glass
    0.96
    blower
    0.96
     玻璃
    0.94
     Bead
    0.93
     shards
    0.91
    fibre
    0.91
    blowing
    0.89
    Activations Density 0.018%

    No Known Activations