© 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. 44495
    Prev
    Next
    INDEX
    Explanations

    court abbreviations (ETB, EEB, DSTB)

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

    The neuron strongly activates on isolated uppercase “B” tokens when they appear as part of legal or academic citation abbreviations (e.g. case reporter or journal acronyms).

    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
    </h5>
    -2.30
    </em>
    -2.06
    </td>
    -1.98
     OGSÅ
    -1.94
    午前
    -1.94
    瑧
    -1.93
    </h4>
    -1.93
    けれ
    -1.74
    </h6>
    -1.73
     Unido
    -1.73
    POSITIVE LOGITS
     deliciously
    1.88
    -),
    1.82
     crucial
    1.82
    B
    1.74
    だったり
    1.70
     lediglich
    1.67
     migh
    1.63
     solamente
    1.58
     essentials
    1.58
     categor
    1.58
    Activations Density 0.079%

    No Known Activations