© 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. 3-RES-JB
    11. 21242
    Prev
    Next
    INDEX
    Explanations

    acronyms related to political or governmental entities

    oai_token-act-pair · gpt-3.5-turbo

    references to the state of Michigan

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    sword
    -0.73
    seeing
    -0.71
    size
    -0.70
    sworth
    -0.70
    space
    -0.70
    kamp
    -0.68
    ãĤ¡
    -0.67
    icular
    -0.67
    comm
    -0.66
     nets
    -0.66
    POSITIVE LOGITS
    ovember
    0.97
    AMI
    0.95
    HAEL
    0.94
    uppet
    0.90
    MI
    0.82
    gow
    0.82
    peg
    0.79
    ascus
    0.78
    OSH
    0.77
    eredith
    0.77
    Activations Density 0.005%

    No Known Activations