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. 8-RES-JB
    6. 8722
    Prev
    Next
    INDEX
    Explanations

    references to the video game "Super Smash Bros", particularly focusing on the character Pikachu

    oai_token-act-pair · gpt-3.5-turbo

    references to the Super Smash Bros. video game series and its characters

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     effic
    -0.79
    igham
    -0.72
     unfocusedRange
    -0.70
    tnc
    -0.70
     accommodation
    -0.68
     amy
    -0.67
     concess
    -0.66
    ointed
    -0.65
     Daniels
    -0.65
     inflamm
    -0.65
    POSITIVE LOGITS
     Wii
    1.12
     Pinball
    1.00
     Metroid
    0.98
    Nintendo
    0.94
     Bros
    0.93
     Kart
    0.92
     Donkey
    0.90
     Capcom
    0.87
     Nintendo
    0.87
     arcade
    0.87
    Activations Density 0.103%

    No Known Activations