© 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. 4-RES-JB
    11. 2408
    Prev
    Next
    INDEX
    Explanations

    references to the metal "brass" or phrases where "brass" is metaphorically used

    oai_token-act-pair · gpt-3.5-turbo

    references to "brass" in various contexts

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

    No Comments

    Negative Logits
    Madison
    -0.77
    nces
    -0.73
    Sit
    -0.70
    Spread
    -0.69
    reddit
    -0.68
    PLIED
    -0.68
    Sher
    -0.67
    ANA
    -0.66
    Tonight
    -0.66
    lying
    -0.65
    POSITIVE LOGITS
     brass
    1.18
    iere
    1.04
    erie
    0.95
     horns
    0.85
    ergic
    0.82
     bells
    0.79
     percussion
    0.78
     bands
    0.78
    undy
    0.76
    tone
    0.76
    Activations Density 0.004%

    No Known Activations