© 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-3-27B-IT
    5. 34-GEMMASCOPE-2-TRANSCODER-262K
    7. 165781
    Prev
    Next
    INDEX
    Explanations

    carefully, gently, delicately

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_34_width_262k_l0_small_affine
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    غو
    0.37
    erView
    0.36
     হৈ
    0.36
    𒊏
    0.35
     deterrent
    0.34
     masculine
    0.34
    孖
    0.33
    CCC
    0.33
    registers
    0.33
    戽
    0.33
    POSITIVE LOGITS
     delicately
    1.14
     carefully
    1.13
     gently
    1.12
     Carefully
    1.09
     careful
    1.07
     осторо
    1.06
     cautiously
    1.03
     Gently
    1.00
     Careful
    1.00
     delicate
    0.98
    Activations Density 0.182%

    No Known Activations