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. 11-RES-JB
    6. 7331
    Prev
    Next
    INDEX
    Explanations

    sentences containing actions related to law enforcement

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "after" that typically reference events or actions following a significant occurrence

    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
    ety
    -0.82
    =-=-=-=-
    -0.81
    meric
    -0.77
    DragonMagazine
    -0.75
    ertain
    -0.75
    zek
    -0.73
    ebook
    -0.72
    fit
    -0.71
    otype
    -0.69
    (*
    -0.69
    POSITIVE LOGITS
     discovering
    1.37
     noticing
    1.27
     receiving
    1.26
     spotting
    1.23
     encountering
    1.22
     refusing
    1.18
     being
    1.14
     escaping
    1.13
     attempting
    1.13
     detecting
    1.13
    Activations Density 0.082%

    No Known Activations