Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 1006
    Prev
    Next
    INDEX
    Explanations

    terms related to medical procedures and specialties, particularly focusing on 'stab', 'venge', and 'nec'

    oai_token-act-pair · gpt-3.5-turbo

    terms associated with violence and revenge

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ãĥ¡
    -0.70
     overflow
    -0.70
     Mellon
    -0.70
    inian
    -0.69
     Hours
    -0.65
    FO
    -0.65
     Airways
    -0.65
     popcorn
    -0.65
     Capacity
    -0.64
    isma
    -0.63
    POSITIVE LOGITS
    stab
    2.81
    venge
    2.33
    nec
    1.66
    sole
    1.36
    rouse
    1.33
     baseman
    1.30
    grave
    1.28
    cler
    1.15
    living
    1.14
    draft
    1.13
    Activations Density 0.065%

    No Known Activations