Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 7-RES-JB
    6. 24314
    Prev
    Next
    INDEX
    Explanations

    expressions of joy and excitement

    oai_token-act-pair · gpt-3.5-turbo

    expressions of enthusiasm or excitement

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    •layer 7 - joy

    No Comments

    Negative Logits
     ÂŃ
    -0.87
    assad
    -0.80
    kefeller
    -0.73
    ÂŃ
    -0.72
     Laden
    -0.68
    ihadi
    -0.62
     Osama
    -0.61
    century
    -0.61
    osate
    -0.61
     Saddam
    -0.60
    POSITIVE LOGITS
     :)
    1.06
     alot
    1.06
     ;)
    1.04
     youtube
    1.01
     Youtube
    0.99
     haha
    0.99
     reddit
    0.98
     devs
    0.97
     pokemon
    0.97
     google
    0.95
    Activations Density 2.006%

    No Known Activations