© 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 EvalsExportsAPI 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. 4-RES-JB
    11. 1945
    Prev
    Next
    INDEX
    Explanations

    the word "except" followed by a description of an exception

    oai_token-act-pair · gpt-3.5-turbo

    phrases that include the word "for" and relevant follow-up terms indicating exceptions or specific conditions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 4-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    4
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    kefeller
    -0.87
    essed
    -0.83
    orthy
    -0.75
    ongyang
    -0.72
    icide
    -0.71
    acements
    -0.70
    gress
    -0.69
    isted
    -0.68
    Registered
    -0.68
    oured
    -0.68
    POSITIVE LOGITS
     caveats
    0.82
     exceptions
    0.81
     occasional
    0.74
     exception
    0.69
     temporarily
    0.67
    ^^^^
    0.66
     mention
    0.65
     accidentally
    0.64
     caveat
    0.64
     fortunately
    0.64
    Activations Density 0.071%

    No Known Activations