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-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13250
    Prev
    Next
    INDEX
    Explanations

    decision-making actions

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

    Voluntary actions and choices

    np_acts-logits-general · gemini-2.0-flash

    making a choice

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 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
     Mastery
    -0.42
     mastery
    -0.40
     assuring
    -0.37
     scores
    -0.37
    NewReader
    -0.36
     rör
    -0.36
    Executor
    -0.35
    Cuáles
    -0.35
     HttpHeaders
    -0.35
    Soli
    -0.34
    POSITIVE LOGITS
     voluntarily
    0.73
     opted
    0.70
     opting
    0.67
     chose
    0.67
    あえて
    0.66
     choose
    0.64
     voluntary
    0.62
     chooses
    0.61
    chose
    0.60
     tercih
    0.60
    Activations Density 0.336%

    No Known Activations