© 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. 0-RES-JB
    11. 20380
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Frag
    -0.72
    oves
    -0.70
    sect
    -0.69
    vid
    -0.68
    cia
    -0.67
    ns
    -0.64
    vi
    -0.63
    agram
    -0.62
     Travels
    -0.62
    Ing
    -0.62
    POSITIVE LOGITS
     answ
    0.86
     hazard
    0.74
     lett
    0.73
    £ı
    0.72
     assurance
    0.70
     indu
    0.70
     dam
    0.67
    ŃĶ
    0.67
    76561
    0.67
     resorts
    0.65
    Activations Density 0.000%

    No Known Activations

    This feature has no known activations.