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. 12271
    Prev
    Next
    INDEX
    Explanations

    verbs that indicate disagreement, approval, or legal terminology

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

    ending in "ing"

    np_acts-logits-general · gemini-2.0-flash

    ing verb endings

    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
     latter
    -0.92
    lanatory
    -0.65
    verwijspagina
    -0.60
     betweenstory
    -0.59
    śmy
    -0.58
    ystems
    -0.58
    bladder
    -0.57
     resourceCulture
    -0.57
    VideoCapture
    -0.56
    pecified
    -0.56
    POSITIVE LOGITS
    ing
    4.21
    ING
    3.10
    ting
    1.76
    ging
    1.57
    izing
    1.53
    ning
    1.47
    ingi
    1.37
    ming
    1.36
    ed
    1.35
    ating
    1.33
    Activations Density 1.099%

    No Known Activations