© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 22-GEMMASCOPE-RES-16K
    6. 12762
    Prev
    Next
    INDEX
    Explanations

    terms and concepts related to mathematical fractions and formats

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

    followed by "format"

    np_acts-logits-general · gemini-2.0-flash

    predicting format after technical terms

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.hook_resid_post
    Hook Layer
    22
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    GHIJKLM
    -0.65
    ✨:
    -0.63
    principalColumn
    -0.62
    igeon
    -0.62
     viewDidLoad
    -0.62
    Personendaten
    -0.61
    ScopeManager
    -0.60
     '+':
    -0.60
    UserScript
    -0.59
    phrine
    -0.56
    POSITIVE LOGITS
     format
    3.60
    format
    3.38
     Format
    3.20
    Format
    3.11
     formats
    2.80
     FORMAT
    2.77
    FORMAT
    2.65
    格式
    2.44
     Formats
    2.42
     formato
    2.38
    Activations Density 0.138%

    No Known Activations