© 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
    2. Gemma-3-27B-IT
    3. 31-GEMMASCOPE-2-TRANSCODER-262K
    4. 166865
    Prev
    Next
    INDEX
    Explanations

    Based on the provided lists, the neuron seems to be associated with programming contexts, specifically with loop counters or dimension variables. The most activating tokens are single letters or common abbreviations like 'n', 'N', or 'SIZE', which are frequently used to denote numerical bounds or counts in code. The tokens that follow these are predominantly programming syntax elements like closing parentheses and semicolons, which typically terminate statements or expressions.Considering the pattern:- `n`, `N`, `SIZE` as variables (often for loops/dimensions).- Punctuation like `):` and `;` following them.- The `TOP_ACTIVATING_TEXTS` show examples like `for i in range(n):`, `for (int i = 0; i < SIZE; i++)`, `for k in range(N):`.These examples strongly

    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_31_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
    Christina
    0.37
     Krist
    0.33
     Christina
    0.33
    Krist
    0.32
    在北京
    0.32
     krist
    0.32
    crystalline
    0.30
     fisk
    0.30
     Кры
    0.30
     एकीकृत
    0.30
    POSITIVE LOGITS
    %(
    0.28
    थ
    0.28
    ="'$(
    0.27
     relacionados
    0.27
    WebV
    0.27
     சை
    0.26
    രു
    0.26
     reorganization
    0.26
     remuneration
    0.26
     nodos
    0.26
    Activations Density 0.001%

    No Known Activations