© 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-3-12B-IT
    3. 12-GEMMASCOPE-2-RES-16K
    4. 7244
    Prev
    Next
    INDEX
    Explanations

    - "her spine straight"- "Why Running Straight Doesn't Work"- "using a straight line (in 2D)"- "flat perspective"- "Flat: Versatile for broad strokes, blending, washes, filling in areas. Can create sharp edges."**Synthesizing the pattern:**The word "straight" appears frequently, often followed by "line" or "lines". It's associated with shapes, movement, and physical configurations.The word "flat" appears, associated with "perspective" and describing surfaces or an art style.The common theme linking "straight" and "flat" in these contexts is geometrical description or perspective.Possible explanations:- straight line or flat perspective- straight geometric descriptions- straight lines and flat perspective- linear and planar descriptions- straight paths and flat surfacesGiven the examples like "straight line", "straight sides", "straight lines", and "flat perspective", the neuron seems to be identifying concepts related to geometry, direction, and visual representation that involve straightness or flatness."straight lines and flat perspective" is 5 words, fits the criteria."straight geometric descriptions" is 3 words."linear and planar descriptions" is 3 words.straight lines and flat perspective

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-it/resid_post/layer_12_width_16k_l0_medium
    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
    েই
    2.66
    ما
    2.56
    ar
    2.45
    лни
    2.36
    و
    2.30
    лда
    2.19
    டன்
    2.14
    ುದು
    2.11
    ют
    2.06
    ství
    2.02
    POSITIVE LOGITS
     demolished
    2.02
    ў
    1.97
    disabled
    1.77
    저
    1.72
     teeth
    1.71
     pollutants
    1.67
     culminated
    1.67
    asında
    1.66
     disintegrated
    1.65
     sheer
    1.62
    Activations Density 0.280%

    No Known Activations