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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 7-RES-JB
    6. 1199
    Prev
    Next
    INDEX
    Explanations

    nutritional information such as daily values and dietary percentages

    oai_token-act-pair · gpt-3.5-turbo

    numerical values and percentages in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    RAW
    -0.67
    GUI
    -0.59
     pseud
    -0.58
    CVE
    -0.58
    WARNING
    -0.58
     Lobby
    -0.56
    pedia
    -0.56
    VPN
    -0.55
    Redditor
    -0.55
    EStream
    -0.54
    POSITIVE LOGITS
    lycer
    0.82
    ngth
    0.76
    ãĥĻ
    0.73
    pees
    0.70
    tein
    0.70
    rated
    0.69
    amic
    0.68
    hya
    0.67
    elsius
    0.67
     polled
    0.66
    Activations Density 0.232%

    No Known Activations