© 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. 19973
    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
    Redditor
    -0.67
    ENA
    -0.64
    achi
    -0.63
    abilia
    -0.63
    Removed
    -0.63
    anta
    -0.62
    anka
    -0.60
     Ancients
    -0.60
    aunts
    -0.58
     unlucky
    -0.58
    POSITIVE LOGITS
    «ĺ
    0.81
    Champ
    0.70
    abies
    0.67
    ocal
    0.66
    ancial
    0.66
    rontal
    0.66
    ohyd
    0.65
    oultry
    0.65
    20439
    0.64
     Steps
    0.62
    Activations Density 0.000%

    No Known Activations

    This feature has no known activations.