© 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. 16328
    Prev
    Next
    INDEX
    Explanations

    information related to transportation or movement of objects

    oai_token-act-pair · gpt-3.5-turbo

    the word "carrying" and its variations related to transportation or conveyance

    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
    ellen
    -0.92
    dash
    -0.71
    vironment
    -0.70
    clair
    -0.66
    debian
    -0.66
    correct
    -0.65
    aligned
    -0.65
    reddits
    -0.64
     Creating
    -0.63
    ior
    -0.63
    POSITIVE LOGITS
     carriers
    0.82
    loads
    0.81
     containers
    0.79
     crates
    0.78
     lantern
    0.76
     Boone
    0.75
     capacity
    0.73
    IGH
    0.71
     suitcase
    0.71
     capacities
    0.71
    Activations Density 0.017%

    No Known Activations