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. 15132
    Prev
    Next
    INDEX
    Explanations

    discussions or concepts related to theoretical frameworks and models

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

    "theory" or "theoretical"

    np_acts-logits-general · gemini-2.0-flash

    theoretical concepts and studies

    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
    ricultural
    -0.47
     Adder
    -0.46
    visits
    -0.44
     garg
    -0.44
     Elvis
    -0.43
     Osm
    -0.43
     cropped
    -0.43
    Musical
    -0.43
    Elvis
    -0.43
     Amar
    -0.42
    POSITIVE LOGITS
     theory
    1.16
     Theory
    1.05
    theory
    1.03
     THEORY
    0.98
    Theory
    0.96
     theories
    0.93
     Theorie
    0.88
     teori
    0.88
     teoría
    0.84
     théorie
    0.83
    Activations Density 0.207%

    No Known Activations