© 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
    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. 5061
    Prev
    Next
    INDEX
    Explanations

    specific names of people, such as celebrities and public figures

    oai_token-act-pair · gpt-3.5-turbo

    references to specific television personalities, particularly late-night show hosts

    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
    dk
    -0.70
    Ë
    -0.70
     GEAR
    -0.70
     hen
    -0.69
    BU
    -0.67
    wings
    -0.67
    MH
    -0.66
     bindings
    -0.66
     wings
    -0.65
     wing
    -0.65
    POSITIVE LOGITS
     Colbert
    3.88
     Limbaugh
    1.37
     Spicer
    1.31
     Fallon
    1.30
     Kimmel
    1.29
     Cosby
    1.17
     Hannity
    1.15
     Constantine
    1.12
     Santorum
    1.12
     Bezos
    1.08
    Activations Density 0.020%

    No Known Activations