© 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. 20933
    Prev
    Next
    INDEX
    Explanations

    phrases indicating a transition or introduction to a new topic

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating a lack of delay or a direct approach to a topic

    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
    eson
    -0.86
    bered
    -0.68
    pak
    -0.66
    payers
    -0.65
    aceous
    -0.63
    tten
    -0.63
    Bird
    -0.61
    nor
    -0.61
     whichever
    -0.60
    ta
    -0.59
    POSITIVE LOGITS
     typo
    0.86
     exaggeration
    0.84
     ado
    0.82
     interruption
    0.77
     hesitation
    0.76
     sarc
    0.72
     understatement
    0.71
     confused
    0.69
     confusion
    0.69
     delay
    0.68
    Activations Density 0.148%

    No Known Activations