Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 12775
    Prev
    Next
    INDEX
    Explanations

    facilities and amenities in a venue.

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    information related to room facilities and rental prices.

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    The neuron appears to be detecting amenities and features of rooms or spaces, particularly in hospitality or real estate contexts. It activates strongly for words and phrases related to facilities, furnishings, and accommodations, such as: - Kitchen appliances and equipment (coffee maker, refrig

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @bot

    Amenities or features of a place

    np_acts-logits-general · gemini-2.0-flash

    items and features present

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    thâu
    -0.94
     "..\..\..\
    -0.75
    soever
    -0.72
    رشف
    -0.70
    хьтан
    -0.68
    ScopeManager
    -0.67
     "..\..\
    -0.66
    machi
    -0.65
     "":
    -0.65
    IndentedString
    -0.65
    POSITIVE LOGITS
     aanwezig
    0.63
     galore
    0.62
     plus
    0.62
     onsite
    0.61
     mevcut
    0.57
    Plus
    0.57
     opcional
    0.56
     completes
    0.54
     included
    0.54
     ausgestattet
    0.54
    Activations Density 0.172%

    No Known Activations