© 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. Gemma-2-2B
    3. 21-GEMMASCOPE-TRANSCODER-16K
    4. 5706
    Prev
    Next
    INDEX
    Explanations

    words that are either very common or signal a transition in an argument

    oai_token-act-pair · gemini-2.0-flash

    forms of the word 'talk'

    np_token-act-pair-logits · gpt-4o-mini

    talk

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ">//
    -0.55
    GOTREF
    -0.53
    XDECREF
    -0.48
    Демографія
    -0.47
    Portail
    -0.45
     juges
    -0.45
    >>();
    -0.43
    TagHelpers
    -0.43
    '];
    
    -0.42
    essoal
    -0.41
    POSITIVE LOGITS
     talk
    2.61
     talks
    2.44
    talk
    2.41
    Talk
    2.36
     Talk
    2.36
     talking
    2.36
     conversation
    2.22
     TALK
    2.22
     talked
    2.22
     conversations
    2.09
    Activations Density 3.773%

    No Known Activations