© 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. 0-RES-JB
    11. 2346
    Prev
    Next
    INDEX
    Explanations

    phrases related to critical points or extreme situations

    oai_token-act-pair · gpt-3.5-turbo

    concepts related to precarious situations or impending crises

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    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
    val
    -0.62
     answered
    -0.61
    var
    -0.60
     Citation
    -0.60
    AM
    -0.59
     properly
    -0.59
    ero
    -0.59
     addressed
    -0.59
    iner
    -0.59
    Met
    -0.57
    POSITIVE LOGITS
     brink
    3.68
     verge
    2.70
     edge
    1.30
     precip
    1.25
     abyss
    1.16
     ledge
    1.07
     outskirts
    1.05
     threshold
    1.02
     eve
    1.01
     borderline
    1.00
    Activations Density 0.015%

    No Known Activations