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. 22036
    Prev
    Next
    INDEX
    Explanations

    phrases related to things being hidden or covered, like buried information or treasure

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of being buried, whether literally or metaphorically

    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
    bal
    -0.83
    ragon
    -0.69
    market
    -0.63
    âĶ
    -0.63
    obe
    -0.62
     Axis
    -0.61
     MEN
    -0.61
    ismo
    -0.61
     âĶĤ
    -0.61
    erc
    -0.61
    POSITIVE LOGITS
     buried
    3.57
     bury
    2.21
     burying
    2.15
     burial
    1.70
     crem
    1.54
     submerged
    1.51
     tucked
    1.39
     hidden
    1.37
     shrouded
    1.28
     drowned
    1.28
    Activations Density 0.018%

    No Known Activations