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. 0-RES-JB
    6. 64
    Prev
    Next
    INDEX
    Explanations

    mentions of sports championships or competition events

    oai_token-act-pair · gpt-3.5-turbo

    references to championships in various sports

    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
     shell
    -0.75
    ola
    -0.62
    ulp
    -0.60
     arrow
    -0.58
     recl
    -0.58
     Maya
    -0.57
     Rabbit
    -0.57
     frightened
    -0.57
    urn
    -0.56
     zombie
    -0.56
    POSITIVE LOGITS
     championships
    2.95
     Championships
    2.75
     Cups
    1.41
     tournaments
    1.40
     competitions
    1.36
     medals
    1.18
    ournaments
    1.10
     races
    1.08
     Races
    1.04
    ampions
    1.01
    Activations Density 0.016%

    No Known Activations