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. 10597
    Prev
    Next
    INDEX
    Explanations

    numbers indicating a quantity or amount, specifically the number 25 with varying degrees of activation

    oai_token-act-pair · gpt-3.5-turbo

    repeated references to the number 25

    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
    •Embedding Space Digit Features

    No Comments

    Negative Logits
    oldemort
    -0.66
     traged
    -0.61
     Gaal
    -0.58
    oké
    -0.58
    aghd
    -0.57
    .""
    -0.57
    erent
    -0.55
     nodd
    -0.55
    cle
    -0.55
    EEE
    -0.55
    POSITIVE LOGITS
     25
    2.91
     26
    2.24
     35
    2.12
     27
    2.12
     30
    2.10
     23
    2.05
     20
    2.03
     24
    2.03
     29
    2.00
     28
    2.00
    Activations Density 0.039%

    No Known Activations