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

    Neuronpedia

    APICircuit 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. 6633
    Prev
    Next
    INDEX
    Explanations

    commands or actions related to engagement and participation.

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

    information about events and activities in a community, including announcements, tours, and concerts.

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

    phrases related to directing attention or taking action, such as "keep an eye out", "check out", "take a look", "give some", etc.

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

    Imperative sentences addressing the reader

    np_acts-logits-general · gemini-2.0-flash

    directs your attention

    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
    WithIOException
    -0.90
    +#+#
    -0.89
    PasswordEncoder
    -0.83
     ElementRef
    -0.75
    Xna
    -0.74
    OGND
    -0.74
     محفوظة
    -0.73
     EconPapers
    -0.73
     linkovi
    -0.72
     GenerationType
    -0.72
    POSITIVE LOGITS
     your
    1.24
     yourself
    1.23
     yourselves
    0.96
     Yourself
    0.89
     Your
    0.83
    your
    0.81
    Your
    0.77
     gärna
    0.71
     yours
    0.70
     me
    0.69
    Activations Density 0.391%

    No Known Activations