© 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 EvalsExportsinterp-engineNEWAPI 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. 12503
    Prev
    Next
    INDEX
    Explanations

    phrases indicating being under some kind of pressure, threat, or investigation

    oai_token-act-pair · gpt-3.5-turbo

    instances where subjects or entities are described as being "under" some form of pressure, threat, or scrutiny

    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
     Begins
    -0.72
    aukee
    -0.69
     hereafter
    -0.62
     Papers
    -0.62
     Retrieved
    -0.58
     Poles
    -0.58
     VIDEOS
    -0.58
     Classics
    -0.58
     mornings
    -0.57
     sake
    -0.56
    POSITIVE LOGITS
    lined
    1.17
    developed
    1.13
    valued
    1.11
    whel
    1.10
    represented
    1.08
    stood
    1.08
    written
    1.07
    powered
    1.05
    paid
    1.05
    lain
    1.05
    Activations Density 0.046%

    No Known Activations