© 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. 8-RES-JB
    11. 17221
    Prev
    Next
    INDEX
    Explanations

    mentions of weight divisions and specific fighters in combat sports

    oai_token-act-pair · gpt-3.5-turbo

    terms related to different weight classes in combat sports, particularly "lightweight" and "heavyweight."

    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
    anooga
    -1.08
    uation
    -0.95
    lon
    -0.92
    lear
    -0.90
    bos
    -0.87
    lee
    -0.87
    ifice
    -0.87
    ibel
    -0.86
    ochet
    -0.86
    ridges
    -0.84
    POSITIVE LOGITS
     denomin
    0.68
     powerhouse
    0.68
     STOR
    0.65
    ETHOD
    0.63
     CTR
    0.63
     magn
    0.62
    ORPG
    0.62
     camer
    0.61
     EDITION
    0.60
     masterpiece
    0.60
    Activations Density 0.117%

    No Known Activations