© 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-3-27B-IT
    5. 19-GEMMASCOPE-2-TRANSCODER-262K
    7. 210265
    Prev
    Next
    INDEX
    Explanations

    Thinking Process:1. **Analyze MAX_ACTIVATING_TOKENS**: The most prominent tokens are "Dwarf" and "Fortress". "Hack" also appears multiple times.2. **Analyze TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: The token "Fortress" frequently follows "Dwarf" or "dwarf". "specific" follows "Hack".3. **Analyze TOP_ACTIVATING_TEXTS**: * "Forgotten Beasts in Dwarf Fortress!" * "Building a Dwarf Fortress-Level of Complexity" * "start of the Dwarf Fortress theme song!Dwarf Fortress, NetHack

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_19_width_262k_l0_small_affine
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     répart
    0.55
     ४
    0.49
     bolas
    0.47
    :=\
    0.47
     четыре
    0.47
     ۴
    0.47
     répartition
    0.47
     myogenic
    0.46
    ắ
    0.46
    ４
    0.46
    POSITIVE LOGITS
     BP
    0.48
    BC
    0.46
     UCS
    0.45
     विनोद
    0.44
     Above
    0.43
    EMS
    0.43
     Zeta
    0.43
     UBC
    0.42
     bp
    0.42
    伯
    0.41
    Activations Density 0.002%

    No Known Activations