© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-2B
    7. Residual Stream - 16k
    9. 18-GEMMASCOPE-RES-16K
    11. 6596
    Prev
    Next
    INDEX
    Explanations

    instances of the word "not" and its variations to identify negations

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

    "Not" followed by an adverb or adjective

    np_acts-logits-general · gemini-2.0-flash

    introducing exceptions or negations

    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
     AssemblyCulture
    -0.95
    ganu
    -0.55
    ]';
    -0.55
     GenerationType
    -0.55
    >>()
    -0.54
     numerus
    -0.54
    Хьажоргаш
    -0.52
     doInBackground
    -0.52
     otomatig
    -0.52
    ыгана
    -0.51
    POSITIVE LOGITS
     quite
    0.77
     everyone
    0.77
     exactly
    0.76
     everybody
    0.73
     sure
    0.71
    withstanding
    0.63
     all
    0.61
    everybody
    0.61
     too
    0.59
     that
    0.59
    Activations Density 0.068%

    No Known Activations