© 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. 7-RES-JB
    11. 12611
    Prev
    Next
    INDEX
    Explanations

    countries or locations

    oai_token-act-pair · gpt-3.5-turbo

    instances of countries or regions mentioned in a geographical context

    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
    iple
    -0.73
    itus
    -0.72
     impulse
    -0.68
     odor
    -0.65
     vibrations
    -0.64
    rive
    -0.64
    raph
    -0.64
    lift
    -0.63
     predec
    -0.62
    ionic
    -0.62
    POSITIVE LOGITS
    England
    0.95
     where
    0.91
    where
    0.88
     Chile
    0.87
     Kazakhstan
    0.86
     Samoa
    0.84
     birthplace
    0.83
    ornia
    0.83
     Territories
    0.83
    Scotland
    0.82
    Activations Density 0.302%

    No Known Activations