© 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. 1725
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical violence and confrontations

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "beating."

    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
    ires
    -0.76
    owan
    -0.71
    alth
    -0.69
    arrow
    -0.66
    uci
    -0.65
    izons
    -0.65
    itional
    -0.65
    Orig
    -0.64
    ague
    -0.63
    endi
    -0.63
    POSITIVE LOGITS
     beating
    3.82
     beat
    2.22
     pounding
    1.95
     beaten
    1.93
     beats
    1.76
    beat
    1.63
     defeating
    1.58
     whipping
    1.57
    Beat
    1.47
     bashing
    1.45
    Activations Density 0.021%

    No Known Activations