© 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. 1-RES-JB
    11. 5529
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ibilities
    -0.78
    Lago
    -0.70
    ifax
    -0.68
     Pinball
    -0.67
    nir
    -0.67
    arin
    -0.67
     UTC
    -0.66
     Leia
    -0.64
     Filip
    -0.64
     Shutterstock
    -0.64
    POSITIVE LOGITS
     eleph
    0.65
     derog
    0.63
     royal
    0.63
    ãĥĨ
    0.62
     innocence
    0.61
     behavi
    0.61
     breat
    0.60
    åİ
    0.59
    éĥ
    0.59
    ologically
    0.59
    Activations Density 0.000%

    No Known Activations

    This feature has no known activations.