© 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 EvalsExportsinterp-engineNEWAPI 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. 11-RES-JB
    11. 5928
    Prev
    Next
    INDEX
    Explanations

    descriptions of sports matches, achievements, and outcomes

    oai_token-act-pair · gpt-3.5-turbo

    terms related to sports events and competitions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Tuls
    -0.96
    olulu
    -0.75
     Anchorage
    -0.72
     Honolulu
    -0.70
    natureconservancy
    -0.68
    Redditor
    -0.66
     Asheville
    -0.65
    Govern
    -0.64
     Rohing
    -0.63
    amaz
    -0.63
    POSITIVE LOGITS
     Champions
    1.55
     UEFA
    1.54
     Europa
    1.50
     Premier
    1.41
     Bundesliga
    1.39
     Premiership
    1.29
     Serie
    1.26
     relegation
    1.25
     Championship
    1.25
     Fifa
    1.20
    Activations Density 0.205%

    No Known Activations