© 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. 8-RES-JB
    11. 17022
    Prev
    Next
    INDEX
    Explanations

    terms related to various types of criminal activities and issues, including murder, rape, violence, corruption, and discrimination

    oai_token-act-pair · gpt-3.5-turbo

    references to violent or criminal acts

    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
    acent
    -0.77
    cill
    -0.76
    herent
    -0.74
    liam
    -0.73
    oodle
    -0.72
    ulum
    -0.70
    cko
    -0.70
    é¾įåĸļ士
    -0.70
    veland
    -0.69
    çīĪ
    -0.68
    POSITIVE LOGITS
     suicides
    1.06
     persecution
    1.03
     alcoholism
    0.97
     harassment
    0.97
     starvation
    0.94
     repression
    0.93
     homelessness
    0.93
     deaths
    0.92
     bloodshed
    0.91
     killings
    0.91
    Activations Density 0.452%

    No Known Activations