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. 710
    Prev
    Next
    INDEX
    Explanations

    parentheses and numbers, especially when used in a legal context

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of parentheses or other bracketed groupings

    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
    utenberg
    -0.65
     hope
    -0.58
    emis
    -0.55
     ÂŃ
    -0.54
    ¬¼
    -0.54
     fears
    -0.53
     tongues
    -0.51
    uca
    -0.51
     fear
    -0.51
     joking
    -0.50
    POSITIVE LOGITS
    (
    3.18
    ((
    2.40
    (-
    2.34
    ([
    2.23
    (*
    2.22
    (&
    2.17
    (_
    2.15
    ("
    2.14
    ($
    2.13
    ({
    1.96
    Activations Density 0.030%

    No Known Activations