© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-2B
    5. 22-RES-MATRYOSHKA-DC
    7. 3148
    Prev
    Next
    INDEX
    Explanations

    phrases like "That being said", "it could be said", and "some might say"

    oai_token-act-pair · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    chanind/gemma-2-2b-batch-topk-matryoshka-saes-w-32k-l0-40/standard/blocks.22.hook_resid_post
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    float32
    Hook Name
    blocks.22.hook_resid_post
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    chanind/pile-uncopyrighted-gemma-1024-abbrv-1B
    Embeds
    IFrame
    Link
    Negative Logits
     notes
    -0.64
    notes
    -0.56
     op
    -0.55
     reminds
    -0.53
     reminded
    -0.53
     explained
    -0.53
     noted
    -0.53
     pointed
    -0.51
     emphasized
    -0.50
    Notes
    -0.49
    POSITIVE LOGITS
     say
    2.34
    say
    2.09
     said
    2.03
    Say
    2.03
     Say
    2.02
     saying
    2.02
     SAY
    1.93
     says
    1.85
     Said
    1.84
    said
    1.80
    Activations Density 0.315%

    No Known Activations