© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 73461
    Prev
    Next
    INDEX
    Explanations

    words preceded by punctuation

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires on the prominent content words of article subtitles or headline sub‐clauses (i.e. the key noun or adjective in the phrase following a colon, dash, or comma in a title).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     exorbit
    -1.32
     horrend
    -1.23
     territor
    -1.15
     chasti
    -1.13
     legendar
    -1.13
     astronom
    -1.11
     indescri
    -1.10
     laug
    -1.10
     qualif
    -1.09
     municip
    -1.08
    POSITIVE LOGITS
     great
    1.28
     big
    1.23
     good
    1.12
     heavy
    1.11
     eternal
    1.05
     open
    1.05
     long
    1.04
     sharp
    1.04
     technological
    1.03
     sweet
    1.02
    Activations Density 0.188%

    No Known Activations