© 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-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 122695
    Prev
    Next
    INDEX
    Explanations

    instances of the word "re" or its variations across different contexts

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

    contractions ending in 're

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    CascadeType
    -0.53
     ElementRef
    -0.49
     möjlighet
    -0.44
     QtGui
    -0.43
    (©
    -0.42
    jaciół
    -0.42
    {{/
    -0.41
    StringUtil
    -0.41
     possibilità
    -0.41
     Schlä
    -0.40
    POSITIVE LOGITS
     theyre
    0.84
     gonna
    0.74
     youre
    0.69
     gettin
    0.56
    s
    0.55
    wiliwch
    0.55
    nosaurus
    0.54
     fueran
    0.53
     goin
    0.53
     been
    0.53
    Activations Density 0.032%

    No Known Activations