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

    Neuronpedia

    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. 2324
    Prev
    Next
    INDEX
    Explanations

    Nouns related to different countries and occupations

    oai_token-act-pair · gpt-3.5-turbo

    references to societal issues and structures

    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
    undai
    -0.88
    bernatorial
    -0.74
    »Ĵ
    -0.71
    anooga
    -0.66
    sym
    -0.62
    tions
    -0.62
    ļé
    -0.61
    assies
    -0.60
    cess
    -0.59
    ¥µ
    -0.58
    POSITIVE LOGITS
     rather
    1.45
     anyway
    1.09
    rather
    1.08
     whereas
    1.04
     anyways
    1.03
     instead
    1.01
    instead
    0.83
     alone
    0.82
    .
    0.80
     regardless
    0.79
    Activations Density 0.774%

    No Known Activations