© 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. 12141
    Prev
    Next
    INDEX
    Explanations

    words related to violence or intense situations

    oai_token-act-pair · gpt-3.5-turbo

    words related to violence or severe injury

    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
     Extensions
    -0.79
     Guides
    -0.69
     Premium
    -0.68
    uci
    -0.67
     Consumer
    -0.65
     Research
    -0.65
     Consumers
    -0.64
    ask
    -0.64
     guides
    -0.64
     Advantage
    -0.63
    POSITIVE LOGITS
     bloody
    3.39
     Bloody
    1.90
     gruesome
    1.60
     bloodshed
    1.53
     murderous
    1.43
     deadly
    1.34
     blood
    1.31
     brutal
    1.31
     carnage
    1.30
     goddamn
    1.30
    Activations Density 0.017%

    No Known Activations