© 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. 18953
    Prev
    Next
    INDEX
    Explanations

    keywords related to physical processes and transformations

    oai_token-act-pair · gpt-3.5-turbo

    references to specific objects or elements in scientific contexts

    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
    Cash
    -0.87
    Faith
    -0.83
    soDeliveryDate
    -0.82
    HUD
    -0.73
    Privacy
    -0.73
    Reddit
    -0.71
    thood
    -0.70
    manship
    -0.70
    FactoryReloaded
    -0.70
    anova
    -0.69
    POSITIVE LOGITS
     slightest
    1.32
     entire
    1.28
     remainder
    1.26
     resultant
    1.22
     outer
    1.21
     wearer
    1.18
     smallest
    1.16
     latter
    1.15
     resulting
    1.15
     entirety
    1.14
    Activations Density 0.488%

    No Known Activations