© 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. 0-RES-JB
    11. 15854
    Prev
    Next
    INDEX
    Explanations

    mentions of objects or items being lit, often with a positive connotation

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "lit" and its variations, primarily in the context of lighting styles in film or photography

    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
     Frazier
    -0.74
     RSS
    -0.67
     Blues
    -0.64
     reb
    -0.63
     assets
    -0.63
    +++
    -0.63
     interstate
    -0.62
     df
    -0.61
     newsletter
    -0.60
     tendencies
    -0.60
    POSITIVE LOGITS
    lit
    4.65
    lighting
    2.00
    light
    1.80
    liter
    1.74
     lit
    1.69
     Lit
    1.63
    lights
    1.35
    lite
    1.25
    Liter
    1.22
    lot
    1.20
    Activations Density 0.008%

    No Known Activations