© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 1-RES-JB
    11. 13963
    Prev
    Next
    INDEX
    Explanations

    references to the business magazine "Fortune"

    oai_token-act-pair · gpt-3.5-turbo

    references to Fortune 500 companies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    liam
    -0.84
    Redditor
    -0.77
    ded
    -0.74
     Hemp
    -0.70
    ream
    -0.68
    ifiers
    -0.68
    ãĤ³
    -0.68
    ounty
    -0.67
    itar
    -0.66
    itations
    -0.65
    POSITIVE LOGITS
     Fortune
    1.10
     tremend
    0.94
     nomine
    0.92
    inarily
    0.76
     contrace
    0.73
     perspect
    0.71
     mosqu
    0.71
     Cookie
    0.70
    DEV
    0.70
    emort
    0.68
    Activations Density 0.005%

    No Known Activations