© 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. 2-RES-JB
    11. 1797
    Prev
    Next
    INDEX
    Explanations

    proper names related to various contexts, potentially including characters, places, and organizations

    oai_token-act-pair · gpt-3.5-turbo

    references to specific characters or personalities, particularly focusing on the name "Kyl."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ername
    -0.79
    ÙĬ
    -0.78
    ãĥ¼ãĥ³
    -0.77
    ä
    -0.74
    Ø©
    -0.73
    recy
    -0.73
    Ùħ
    -0.72
    ãĥīãĥ©
    -0.71
    EMS
    -0.70
    代
    -0.70
    POSITIVE LOGITS
    bone
    0.82
    alos
    0.81
    hawk
    0.76
    sten
    0.76
    ston
    0.74
     Luthor
    0.73
    omore
    0.72
    Chel
    0.72
     Kang
    0.70
     lapt
    0.70
    Activations Density 0.007%

    No Known Activations