Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 13082
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate coercion or compulsion

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

    "forced" and its variations

    np_acts-logits-general · gemini-2.0-flash

    forced to act

    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
     Komunikasi
    -0.50
     amarillas
    -0.47
     hva
    -0.47
     cielos
    -0.46
    crí
    -0.44
     quedado
    -0.43
     rubia
    -0.42
     Vorbe
    -0.42
     Desarrollo
    -0.41
     embaj
    -0.41
    POSITIVE LOGITS
    forced
    1.13
     Forced
    1.10
    Forced
    1.03
     forced
    1.02
     Compulsory
    0.89
     compulsory
    0.82
    被迫
    0.80
    ulsory
    0.80
     Force
    0.77
     forcing
    0.77
    Activations Density 0.275%

    No Known Activations