© 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. 10-RES-JB
    11. 12065
    Prev
    Next
    INDEX
    Explanations

    mentions of the country "Saudi Arabia" in various contexts

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    mble
    -0.85
    early
    -0.71
    ombie
    -0.71
    lessly
    -0.68
    ucl
    -0.66
    alach
    -0.66
    lr
    -0.66
    obyl
    -0.65
    ombies
    -0.65
    sc
    -0.64
    POSITIVE LOGITS
     Arabia
    2.27
     Arabian
    1.84
     Aram
    1.38
     Riy
    0.98
     Saud
    0.98
     Abdullah
    0.97
     Abdul
    0.95
     Arab
    0.95
     Riyadh
    0.93
     princes
    0.92
    Activations Density 0.017%

    No Known Activations