© 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. 11-RES-JB
    11. 5739
    Prev
    Next
    INDEX
    Explanations

    concepts related to physical violence and assault

    oai_token-act-pair · gpt-3.5-turbo

    instances of violence and abuse

    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
    20439
    -0.67
     Principle
    -0.63
     CFR
    -0.61
    aird
    -0.61
     understatement
    -0.60
     Depot
    -0.59
     Focus
    -0.59
     Rae
    -0.58
    ahime
    -0.58
     inherited
    -0.57
    POSITIVE LOGITS
     by
    1.23
     merciless
    1.14
     verbally
    1.04
     BY
    1.01
     violently
    1.00
     repeatedly
    0.96
     senseless
    0.95
    by
    0.94
     brutally
    0.92
     unconscious
    0.88
    Activations Density 0.188%

    No Known Activations