© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 487
    Prev
    Next
    INDEX
    Explanations

    instances of harassment, including verbal harassment, physical violence, threats, and intimidation

    oai_token-act-pair · gpt-3.5-turbo

    references to harassment and associated threats or violence

    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
    éĹĺ
    -0.87
    ACTED
    -0.76
     Wonders
    -0.72
    cise
    -0.72
    urgy
    -0.69
    utonium
    -0.68
    uncture
    -0.68
    ederal
    -0.66
    bits
    -0.66
    ECD
    -0.65
    POSITIVE LOGITS
     harassing
    1.03
     tactics
    1.02
     harassment
    0.96
     harass
    0.95
     hurled
    0.92
     campaigns
    0.92
     stalking
    0.88
     accus
    0.88
    assment
    0.85
     leveled
    0.85
    Activations Density 0.104%

    No Known Activations