© 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. 22-GEMMASCOPE-RES-131K
    7. 35421
    Prev
    Next
    INDEX
    Explanations

    city selection practice

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

    The neuron strongly activates on occurrences of technical directional adverbs (especially “downstream” and similar flow‐direction words).

    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
     There
    -1.05
    –
    -0.87
    őre
    -0.87
    的他
    -0.81
    ázaro
    -0.81
     площадь
    -0.78
    We
    -0.77
     его
    -0.77
    しながら
    -0.76
    While
    -0.76
    POSITIVE LOGITS
     recent
    0.93
     raso
    0.91
    рино
    0.90
    isés
    0.89
     ẞ
    0.87
    íble
    0.84
    nesc
    0.82
    yadh
    0.82
    valente
    0.80
     MUNICIP
    0.79
    Activations Density 0.765%

    No Known Activations