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. 7728
    Prev
    Next
    INDEX
    Explanations

    operations related to data input/output functions and processes

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

    Saving, creating, reading, or updating records/files

    np_acts-logits-general · gemini-2.0-flash

    action verbs: save, read, create, update, write, speak

    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
     of
    -0.52
     we
    -0.49
     am
    -0.47
     co
    -0.43
     ar
    -0.42
     bar
    -0.42
     
    -0.42
    <eos>
    -0.42
    ArgsConstructor
    -0.40
     Smith
    -0.38
    POSITIVE LOGITS
     myſelf
    0.84
     ſeveral
    0.80
     Efq
    0.78
     XCTest
    0.76
    ειτουργ
    0.75
     themſelves
    0.73
     pleaſure
    0.73
     itſelf
    0.72
     Eſ
    0.72
     numberOfRows
    0.71
    Activations Density 0.298%

    No Known Activations