© 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. 7472
    Prev
    Next
    INDEX
    Explanations

    * "**Option 3 (Focus on STEM):** [Your Name] will graduate from [High School Name] where they were a key member of the Robotics"* "**Option 1: Formal & Detailed (Good for STEM, Quantitative Fields)** ```email Subject: Research Opportunity Inquiry - [Your Name] - [Course Name]```"* "**Machine Learning (ML) and Deep Learning:** I can utilize pre-trained models (like BERT, RoBERTa or similar transformer-based models) fine-tuned for ABSA tasks. These models have learned patterns in language"* "what's the difference between ARM and RISC-V?"* "order of operations (PEMDAS/BODMAS)"* "order of operations (PEMDAS/BODMAS)"* "Service to Professional Societies: Working within professional organizations (e.g., IEEE, ACM, APA).Based on the provided lists, the neuron seems to be activated by terms related to specific fields of study, abbreviations, and organizational acronyms.Let's break it down:* **MAX_ACTIVATING_TOKENS**: Contains "STEM", "PEM" (likely from PEMDAS), "RIS" (likely from RISC-V), and "IEEE".* **TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: Shows context like "Education", "DAS" (could be part of acronyms), and commas/parentheses for separation.* **TOP_POSITIVE_LOGITS**: Includes suffixes and partial words that might follow specific terms (e.g., "tion", "ically", "less").* **TOP_ACTIVATING_TEXTS**: Explicitly mentions "STEM Education", "Robotics" (related to STEM), "BERT" (ML model), "RISC-V", "PEMDAS", and "IEEE".The common theme is technical, academic, or standardized abbreviations and terms.The most specific and recurring pattern pointing to a concept is related to education and technical acronyms/abbreviations.**Explanation:** STEM, PEMDAS, IEEE, RISC-V

    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
    TARGET
    1.28
    តែ
    1.22
    ’
    1.20
     인한
    1.19
    할
    1.16
    কে
    1.15
    ‌ی
    1.15
     sails
    1.13
     dogged
    1.13
     Архиви
    1.11
    POSITIVE LOGITS
    tedir
    2.22
    oretically
    2.03
    lardan
    1.95
    theless
    1.91
    ו
    1.81
    ação
    1.78
    lere
    1.77
    larda
    1.67
    iation
    1.65
    list
    1.63
    Activations Density 0.080%

    No Known Activations