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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 2-RES-JB
    11. 23148
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    lamm
    -0.68
    rations
    -0.65
    finished
    -0.65
    bub
    -0.65
    strate
    -0.62
    anian
    -0.62
    rawl
    -0.61
    undred
    -0.61
    ubb
    -0.61
    hetics
    -0.61
    POSITIVE LOGITS
     paternity
    0.77
     Clarks
    0.66
    amen
    0.64
     closure
    0.63
     Canaver
    0.63
    izens
    0.62
     Sinn
    0.61
    ICA
    0.61
    76561
    0.60
     paran
    0.59
    Activations Density 0.000%

    No Known Activations

    This feature has no known activations.