© 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 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. 8-RES-JB
    11. 10244
    Prev
    Next
    INDEX
    Explanations

    words related to negative behaviors or actions, particularly harassment and violence

    oai_token-act-pair · gpt-3.5-turbo

    references to harassment and related negative experiences

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    éĹĺ
    -1.03
    iets
    -0.78
    icts
    -0.77
    ethe
    -0.72
    essential
    -0.72
    zyme
    -0.71
    stanbul
    -0.70
    rient
    -0.70
    ernels
    -0.69
    arch
    -0.69
    POSITIVE LOGITS
     accus
    0.92
     harassment
    0.88
     harassing
    0.88
     harass
    0.87
     tactics
    0.84
    assment
    0.82
     stalking
    0.79
     complaints
    0.78
     leveled
    0.77
     allegations
    0.77
    Activations Density 0.060%

    No Known Activations