© 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
    2. Gemma-3-27B-IT
    3. 11-GEMMASCOPE-2-TRANSCODER-262K
    4. 93405
    Prev
    Next
    INDEX
    Explanations

    The user wants an explanation of a neuron's behavior based on the provided lists.The explanation needs to be 3-20 words, a phrase, and specific.I should not use phrases like "This neuron detects/predicts".I should look for patterns in `MAX_ACTIVATING_TOKENS`, `TOKENS_AFTER_MAX_ACTIVATING_TOKEN`, `TOP_POSITIVE_LOGITS`, and `TOP_ACTIVATING_TEXTS`.Let's break down the input:**MAX_ACTIVATING_TOKENS**:- ida- ida- ellig- ellig- fearless- fearless- ionate- fearless- uso- Jelly**TOKENS_AFTER_MAX_ACTIVATING_TOKEN**:- ",- heid- heid- experimental- explorer- .- commitment- con- fin**TOP_POSITIVE_LOGITS**:- ас- টর- akanan- сцене- тор- ту- ční- änä- Teste- bukanlah**TOP_ACTIVATING_TEXTS**:- Pregnancy Terms:** "pregnant", "pregnancy", "gestation", "antenatal", "postpartum", "gravida", "para", "fetal", "fetus", "amniotic fluid", "placenta", "umb- Profile Summary:** * **Age:** 30 years old * **Obstetrical History:** Gravida 1, Para 1 (meaning she's been pregnant once and delivered one baby) * **Medical- ictions" about money (like offering to pay when you don't really intend to). * **"Gezelligheid" doesn't necessarily equal extravagance:** *Gezelligheid* is a core Dutch concept – a feeling of- . This captivating city is a living, breathing tapestry woven with history, art, innovation, and a uniquely Dutch *gezelligheid* (a feeling of cozy conviviality). Whether you're a first-time visitor or a seasoned traveler- Percussion-Heavy) * **Why:** Merrill Garbus (Tune-Yards) is a fearless experimentalist, but at her core she makes incredibly catchy and emotionally powerful music.jellyfin

    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_11_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
     Anchorage
    0.56
    ]$
    0.53
     Ar
    0.52
     Amtrak
    0.52
     Peyton
    0.49
     Designers
    0.49
     Yosemite
    0.48
     Duan
    0.48
     Orientation
    0.48
     Clearing
    0.47
    POSITIVE LOGITS
    ас
    0.54
    ટર
    0.47
    akanan
    0.47
     сцене
    0.44
    тор
    0.43
    ту
    0.43
    ční
    0.43
    änä
    0.43
    Teste
    0.43
     bukanlah
    0.43
    Activations Density 0.000%

    No Known Activations