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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 3-RES-JB
    11. 15747
    Prev
    Next
    INDEX
    Explanations

    sporting events and competitions

    oai_token-act-pair · gpt-3.5-turbo

    references 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 @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     informants
    -0.69
    reprene
    -0.64
     popul
    -0.63
     coerc
    -0.62
     synt
    -0.62
     answ
    -0.61
    weak
    -0.60
    eret
    -0.59
     informant
    -0.58
     predec
    -0.58
    POSITIVE LOGITS
     Finals
    0.86
     Championships
    0.80
    Tickets
    0.80
     Coliseum
    0.78
     SPORTS
    0.76
     2019
    0.75
     Season
    0.74
     Championship
    0.74
     Playoffs
    0.74
     Champions
    0.74
    Activations Density 1.560%

    No Known Activations