© 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 EvalsExportsAPI 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. 8-RES-JB
    11. 10772
    Prev
    Next
    INDEX
    Explanations

    phrases related to concealing or covering something physically

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "cover" and its related forms, particularly in the context of faces and concealment

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Enh
    -0.69
    nir
    -0.68
    Magikarp
    -0.64
    friend
    -0.64
    esson
    -0.64
    Lear
    -0.63
    Cop
    -0.62
    elected
    -0.61
    Member
    -0.61
    Eng
    -0.61
    POSITIVE LOGITS
     bases
    1.15
     entirety
    1.02
     gaps
    0.93
    phia
    0.88
     entire
    0.81
     width
    0.80
    topic
    0.80
     tracks
    0.80
     gap
    0.79
     breadth
    0.79
    Activations Density 0.137%

    No Known Activations