© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 102403
    Prev
    Next
    INDEX
    Explanations

    Arabidopsis, traceback, Faulkner

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

    scientific or technical names, particularly plant/organism names like "Arabidopsis" and programming references like "traceback".

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron flags rare or unusual capitalized tokens—primarily proper names and specific named entities.

    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
    vantaged
    -2.53
    ?"
    -2.44
    </h3>
    -2.39
    िंग
    -2.39
    י
    -2.34
    ي
    -2.34
     While
    -2.23
     Many
    -2.14
    </em>
    -2.09
     When
    -2.09
    POSITIVE LOGITS
     largely
    2.56
     potem
    2.48
     使っ
    2.42
    悻
    2.42
     hugely
    2.42
    たまに
    2.42
     renowned
    2.41
    ſſen
    2.36
    </b>
    2.36
     trois
    2.30
    Activations Density 0.023%

    No Known Activations