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. 18-GEMMASCOPE-RES-16K
    6. 8534
    Prev
    Next
    INDEX
    Explanations

    the recurring use of the word "always" in various contexts

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

    "Always" followed by an adjective or verb

    np_acts-logits-general · gemini-2.0-flash

    always + [verb/adjective]

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    出版年
    -0.76
    er
    -0.67
    ه
    -0.66
     IDisposable
    -0.62
    RuleContext
    -0.59
     Heathrow
    -0.58
     GenerationType
    -0.57
    Niels
    -0.56
     UVB
    -0.56
     Huddersfield
    -0.55
    POSITIVE LOGITS
     luôn
    0.73
     surla
    0.71
    しも
    0.68
     ALWAYS
    0.67
    ALWAYS
    0.66
    Always
    0.63
    IVEREF
    0.63
     Always
    0.61
     Kết
    0.60
    Aims
    0.59
    Activations Density 0.051%

    No Known Activations