© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 3-RES-JB
    11. 18246
    Prev
    Next
    INDEX
    Explanations

    words related to Microsoft Surface products

    oai_token-act-pair · gpt-3.5-turbo

    references to the Microsoft Surface product line

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ais
    -0.86
    sen
    -0.85
    orn
    -0.82
    eros
    -0.81
    atched
    -0.79
    oning
    -0.79
    hovah
    -0.79
    atana
    -0.79
    tar
    -0.75
    ard
    -0.75
    POSITIVE LOGITS
     Surface
    1.04
    Pad
    0.78
     Dock
    0.77
     tablets
    0.76
     Air
    0.75
     Cortex
    0.73
    pad
    0.73
     Gum
    0.73
    pring
    0.70
     Sensor
    0.70
    Activations Density 0.006%

    No Known Activations