© 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. 24360
    Prev
    Next
    INDEX
    Explanations

    references to historical events involving the Nazi party

    oai_token-act-pair · gpt-3.5-turbo

    references to the Nazi regime and related terms

    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
    tis
    -0.90
    pring
    -0.88
    trak
    -0.85
    Downloadha
    -0.79
    area
    -0.77
    region
    -0.77
    ately
    -0.76
    Score
    -0.73
    Asia
    -0.72
    20439
    -0.72
    POSITIVE LOGITS
     sympath
    0.88
     paramilitary
    0.85
     salute
    0.84
     propaganda
    0.83
     Youth
    0.80
     Holocaust
    0.80
     Reich
    0.80
     dictator
    0.79
     occupation
    0.78
     collaborator
    0.78
    Activations Density 0.011%

    No Known Activations