© 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
    3. 31-GEMMASCOPE-2-RES-262K
    4. 174897
    Prev
    Next
    INDEX
    Explanations

    full moon party

    np_acts-logits-general · gemini-2.5-flash-lite

    specialty foodlifeguard hurriedinfinite detailsprofiles forSpecialty Food StoresProfiles of buyersspecialty foodlifeguardSpecialty Food StoreslifeguardProfiles for buyersspecialty foodProfiles of buyersProfiles of buyersProfiles of buyersProfiles of buyersProfiles of buyersProfiles of buyersProfiles of buyersProfiles of buyersProfiles of buyers</TOP_POSITIVE_LOGITS><TOP_ACTIVATING_TEXTS> and wind) has plummeted in recent years, making them increasingly competitive with fossil fuels. * **Reliability/Intermittency. It provides data structures like DataFrames, which are like spreadsheets. We use the alias `pd`. * **`import matplotlib.pyplot as plt`**: This imports the `pyplot` module from the Matplotlib library. whole, at different magnifications. It's not *exactly* the same, but statistically similar. * **Infinite Detail:** You can theoretically zoom in on a fractal forever and continue to find new, intricate details. ThereUsually a couple of options, often store brands or less well-known brands. Quality can vary. **2. Specialty Food Stores** * **Olive Oil Shops/Tasting Rooms:** These are becoming more common, especially in, with a sinking feeling, she realized *more* than that was missing. Her face burned with embarrassment as a lifeguard hurried over with a towel and a concerned expression. She wrapped the towel tightly around herself, trying to disappear. in water, streamlined body, natural weaponry (teeth), senses geared towards underwater combat (electroreception). * **Weaknesses:** Completely helpless on land. Limited maneuverability outside of water. Relatively fragile outside of its protectiveatic agents! It's a key characteristic that contributes to their effectiveness. Here's a breakdown of what "fluidity" means in this context, going beyond just "flows easily," and exploring the factors influencing it: **What does is a simplification:** It doesn't fully account for grid integration costs (very important for renewables), intermittency, or the value of energy at different times of day. * **Rapidly Changing Costs:** Renewable energy costs, breakdown of how AI can be used, categorized by function, with examples and tools: **I. Identifying Ideal Customer Profiles (ICPs) & Building Target Lists** * **AI-Powered Data Enrichment:** AI can analyze existing customer **Benefits of Cloud Desktops (DaaS) over Local Desktops:** | Feature | Cloud Desktop (DaaS) | Local Desktop | |---|---|---| | **Cost** | OPEX (Operational Expenditure) -</TOP_ACTIVATING_TEXTS>**MAX_ACTIVATING_TOKENS**:mittSpecialtylifeguarditytheProfilesaaS**TOKENS_AFTER_MAX_ACTIVATING_TOKEN**:ency (follows 'ity')**Infinite (follows 'Infinite')Food (follows 'Specialty')hurried (follows 'lifeguard')**"value (follows 'value')(**TOP_POSITIVE_LOGITS**:femalecomplainingrelievingflushingimperfectionjarringslammingོhomebuyerscramping**TOP_ACTIVATING_TEXTS**:- "Specialty Food Stores"- "lifeguard hurried"- "Infinite Detail"- "Profiles (ICPs) & Building Target Lists"- "DaaS (Desktop as a Service)" - related to 'aaS'Let's look for patterns:1. **MAX_ACTIVATING_TOKENS**: 'Specialty', 'lifeguard', 'profiles', 'aaS'. These seem like specific nouns or technical terms.2. **TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: 'Food' (after 'Specialty'), 'hurried' (after 'lifeguard'), 'Infinite' (after 'Infinite'), 'value' (after 'value'). customer profiles

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-pt/resid_post/layer_31_width_262k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     CONDUCT
    0.47
    葷
    0.47
    枧
    0.47
    注释
    0.45
     মন্তব্য
    0.44
     তন্ম
    0.43
    ত্ম
    0.42
    ധാന
    0.42
    స్తాయి
    0.41
    zhou
    0.41
    POSITIVE LOGITS
     female
    0.52
     complaining
    0.49
     femenina
    0.49
     murdering
    0.47
    ོ
    0.46
     femme
    0.46
     flushing
    0.46
     homebuyers
    0.46
    female
    0.46
     guessing
    0.46
    Activations Density 0.000%

    No Known Activations