© 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. 4-RES-JB
    11. 6700
    Prev
    Next
    INDEX
    Explanations

    mentions of the city "Dubai"

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the location "Dubai."

    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
    otin
    -0.83
    ym
    -0.79
    avery
    -0.77
    baugh
    -0.72
    ages
    -0.72
    urgical
    -0.70
    о
    -0.70
    avement
    -0.69
    urgy
    -0.67
    essee
    -0.66
    POSITIVE LOGITS
    Leaks
    0.93
     Dhabi
    0.91
     Dubai
    0.89
     Municip
    0.87
     Cooperation
    0.81
    querque
    0.80
     Emir
    0.77
     Lumpur
    0.76
     Marina
    0.74
     Harbour
    0.69
    Activations Density 0.006%

    No Known Activations