© 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. 19210
    Prev
    Next
    INDEX
    Explanations

    specific locations or geographic areas, specifically with the word "Heights" in them

    oai_token-act-pair · gpt-3.5-turbo

    references to specific geographical locations, particularly those with "Heights" in their name

    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
     logging
    -0.76
     repl
    -0.74
    UF
    -0.68
     rune
    -0.67
     stamp
    -0.66
    isu
    -0.65
     rubber
    -0.64
     replication
    -0.62
     reflection
    -0.62
     referee
    -0.61
    POSITIVE LOGITS
     Heights
    4.39
     heights
    1.46
     Township
    1.08
    sburg
    1.02
    atown
    0.98
     Hills
    0.97
    hattan
    0.95
     Neighborhood
    0.94
     Cemetery
    0.92
     Borough
    0.90
    Activations Density 0.009%

    No Known Activations