© 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. 22-GEMMASCOPE-RES-131K
    7. 70697
    Prev
    Next
    INDEX
    Explanations

    missed opportunity

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

    This neuron isn’t detecting any linguistic pattern or meaning but simply emits a value proportional to each token’s numeric ID (i.e. it acts as a “vocabulary‐index” passthrough).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
    твер
    -0.76
    placeholder
    -0.73
     !=
    -0.73
    lineNumber
    -0.73
    něte
    -0.73
    itation
    -0.72
    breakpoints
    -0.71
     Pertama
    -0.71
    щении
    -0.70
     occuper
    -0.70
    POSITIVE LOGITS
     missed
    1.70
     could
    1.24
     but
    1.23
     Missed
    1.20
    Missed
    1.18
    missed
    1.11
     hätte
    1.05
    惜
    1.03
     timing
    1.02
     tiế
    1.02
    Activations Density 0.041%

    No Known Activations