© 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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 6612
    Prev
    Next
    INDEX
    Explanations

    phrases and concepts related to long-term relationships and commitments

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

    long duration

    np_acts-logits-general · gemini-2.0-flash

    long-term lasting

    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
     المعيارى
    -0.47
    msgTypes
    -0.41
     string
    -0.38
     disemb
    -0.37
    endientes
    -0.36
     tiet
    -0.36
     Holiness
    -0.35
     الاخت
    -0.35
     parah
    -0.35
     Morg
    -0.35
    POSITIVE LOGITS
    forever
    0.74
     longevity
    0.72
     forever
    0.71
     perpetuity
    0.67
    durable
    0.63
     durable
    0.62
    Forever
    0.61
     permanence
    0.60
     FOREVER
    0.60
     lasting
    0.59
    Activations Density 0.378%

    No Known Activations