© 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. 6999
    Prev
    Next
    INDEX
    Explanations

    pairs of individuals in official or leadership positions

    oai_token-act-pair · gpt-3.5-turbo

    conjunctions and references to groups of individuals or roles in a context

    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
    aves
    -0.82
    imil
    -0.80
    mop
    -0.77
    blocks
    -0.77
    raq
    -0.77
    iem
    -0.74
    aea
    -0.74
    aires
    -0.73
    atten
    -0.72
    aven
    -0.71
    POSITIVE LOGITS
     others
    0.93
     former
    0.92
     longtime
    0.89
     deputy
    0.88
     drummer
    0.86
     colleagues
    0.86
     Vice
    0.85
     Frenchman
    0.85
     novelist
    0.84
     assorted
    0.83
    Activations Density 0.152%

    No Known Activations