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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 0-RES-JB
    11. 6163
    Prev
    Next
    INDEX
    Explanations

    references to a good sense of humor in text

    oai_token-act-pair · gpt-3.5-turbo

    references to humor

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Tata
    -0.77
     Transfer
    -0.74
    Transfer
    -0.74
     Isis
    -0.71
    izu
    -0.69
    FO
    -0.67
    DA
    -0.67
    MP
    -0.66
     Awoken
    -0.65
     NSW
    -0.65
    POSITIVE LOGITS
     humor
    3.95
     humour
    3.12
     humorous
    1.88
     jokes
    1.83
     satire
    1.79
     comedy
    1.70
     comedic
    1.68
     satir
    1.68
     laughter
    1.60
     joke
    1.56
    Activations Density 0.016%

    No Known Activations