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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 16307
    Prev
    Next
    INDEX
    Explanations

    formatting specifications and patterns for date and time representations

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

    formatting, especially dates and numbers

    np_acts-logits-general · gemini-2.0-flash

    formatting strings

    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
    openqa
    -0.59
    ioutil
    -0.55
     minOccurs
    -0.50
     crystall
    -0.49
     Xunit
    -0.48
    migrationBuilder
    -0.46
     dubbio
    -0.45
     étudier
    -0.44
    orElseThrow
    -0.44
    ább
    -0.44
    POSITIVE LOGITS
     format
    1.29
    format
    1.21
     Format
    1.20
     formatting
    1.13
     formats
    1.11
    Format
    1.10
     FORMAT
    1.07
     formatted
    1.05
     formatter
    1.03
    FORMAT
    1.01
    Activations Density 0.344%

    No Known Activations