© 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. 3081
    Prev
    Next
    INDEX
    Explanations

    words related to aerial activities, such as "aerial," "airborne," and "surveillance."

    oai_token-act-pair · gpt-3.5-turbo
    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
    士
    -0.95
    lda
    -0.87
    shire
    -0.82
    hof
    -0.80
    bard
    -0.79
    urion
    -0.77
    ublic
    -0.76
    pton
    -0.75
    esson
    -0.74
    hoff
    -0.74
    POSITIVE LOGITS
     refuel
    1.04
     aerial
    1.02
     bombardment
    0.97
    plane
    0.97
     reconnaissance
    0.96
     spraying
    0.89
     photography
    0.82
     inspection
    0.80
    ists
    0.80
     photographs
    0.79
    Activations Density 0.023%

    No Known Activations