© 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. 9268
    Prev
    Next
    INDEX
    Explanations

    words related to specific geographic locations, especially those with names containing unusual characters

    oai_token-act-pair · gpt-3.5-turbo

    geographical locations, particularly those related to the Suez Canal and nearby areas

    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
    requ
    -0.63
    tackle
    -0.63
     WH
    -0.63
     captcha
    -0.61
    orem
    -0.61
     AFTER
    -0.58
    requires
    -0.56
    nr
    -0.56
    =-=-=-=-
    -0.55
    govtrack
    -0.55
    POSITIVE LOGITS
    Corp
    0.82
    Plex
    0.81
     Lumpur
    0.76
    Cola
    0.75
    Bank
    0.72
     Esports
    0.70
    ignt
    0.67
     Depot
    0.67
    Sport
    0.66
    pload
    0.66
    Activations Density 0.210%

    No Known Activations