© 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
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 4-RES-JB
    6. 16235
    Prev
    Next
    INDEX
    Explanations

    phrases describing threats or actions of potential harm towards someone or something

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating threats or intentions to harm

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

    No Comments

    Negative Logits
    CTV
    -0.77
     guiName
    -0.71
    ritz
    -0.68
    unny
    -0.68
    iencies
    -0.66
    ROM
    -0.63
    mont
    -0.62
     Tunnel
    -0.62
     Bought
    -0.62
    DragonMagazine
    -0.62
    POSITIVE LOGITS
     retribution
    1.00
     retaliation
    0.94
     harming
    0.76
     vengeance
    0.75
     boycott
    0.71
     retaliate
    0.71
     repr
    0.69
     annihilation
    0.68
     revenge
    0.67
     Juda
    0.67
    Activations Density 0.091%

    No Known Activations