© 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. 22504
    Prev
    Next
    INDEX
    Explanations

    descriptions of criminal activities or events

    oai_token-act-pair · gpt-3.5-turbo

    references to criminal activities and incidents involving victims

    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
    Grade
    -0.71
     ®
    -0.68
    ivably
    -0.68
    !]
    -0.67
    erning
    -0.67
    2020
    -0.67
    folio
    -0.66
    ularity
    -0.64
    !)
    -0.64
     Prediction
    -0.64
    POSITIVE LOGITS
     orally
    0.99
     smelled
    0.91
     mol
    0.89
     Cosby
    0.86
     "'
    0.84
     intoxicated
    0.83
     verbally
    0.82
     panicked
    0.82
     abused
    0.82
     raped
    0.81
    Activations Density 0.572%

    No Known Activations