© 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. 11-RES-JB
    11. 1609
    Prev
    Next
    INDEX
    Explanations

    physical violence acts like hitting or punching

    oai_token-act-pair · gpt-3.5-turbo

    references to injuries or attacks to the body

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    istries
    -0.86
    urable
    -0.82
    nces
    -0.74
    itives
    -0.73
    ét
    -0.69
    Write
    -0.64
    stall
    -0.64
    erion
    -0.63
    selling
    -0.63
    deals
    -0.62
    POSITIVE LOGITS
     groin
    1.53
     abdomen
    1.51
     stomach
    1.42
     chest
    1.41
     forehead
    1.40
     crotch
    1.39
     throat
    1.35
     genitals
    1.35
     thigh
    1.35
     buttocks
    1.35
    Activations Density 0.061%

    No Known Activations