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. 391
    Prev
    Next
    INDEX
    Explanations

    mentions of efficiency and effectiveness in contexts relating to processes or systems

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

    Adjective followed by "manner" or "fashion."

    np_acts-logits-general · gemini-2.0-flash

    adverb preceding manner

    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
    •honest

    No Comments

    Negative Logits
    enumii
    -0.75
    enumi
    -0.73
     للمعارف
    -0.65
     surla
    -0.65
    出版年
    -0.64
     Winaray
    -0.63
    mybatisplus
    -0.61
     betweenstory
    -0.61
    ագրություններ
    -0.60
    Rujuakan
    -0.59
    POSITIVE LOGITS
     manner
    1.86
     fashion
    1.54
    manner
    1.48
     Manner
    1.35
    fashion
    1.24
     maniera
    1.17
     way
    1.12
     manière
    1.07
     façon
    1.03
     style
    1.03
    Activations Density 0.402%

    No Known Activations