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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 11-RES-JB
    11. 21076
    Prev
    Next
    INDEX
    Explanations

    terms related to exploration and investigation

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "explore" and related terms indicating investigation or examination

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    appointed
    -0.78
    warning
    -0.71
    gage
    -0.69
    cour
    -0.67
    Torrent
    -0.66
    corn
    -0.66
    ndra
    -0.65
    fin
    -0.65
    nery
    -0.63
    Ĥİ
    -0.61
    POSITIVE LOGITS
     whether
    1.02
     ways
    1.01
     how
    0.99
     why
    0.96
     possibilities
    0.95
     topics
    0.92
     WHY
    0.90
     alternatives
    0.85
    gdala
    0.82
     themes
    0.82
    Activations Density 0.142%

    No Known Activations