© 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
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 7-RES-JB
    6. 17231
    Prev
    Next
    INDEX
    Explanations

    the word "president" followed by an organization or title

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of presidency and is primarily focused on the phrase "president of the United States"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     descriptions
    -0.81
     dishes
    -0.74
     breeze
    -0.73
     deficiencies
    -0.72
     informants
    -0.72
     dramas
    -0.70
     dish
    -0.70
     needles
    -0.69
     relate
    -0.68
     setbacks
    -0.67
    POSITIVE LOGITS
     Nations
    0.82
     United
    0.79
     Austral
    0.79
     UNESCO
    0.78
     Worldwide
    0.77
     Liberty
    0.76
     UNITED
    0.76
     Association
    0.75
     Organization
    0.75
     Conference
    0.75
    Activations Density 0.074%

    No Known Activations