© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 4-RES-JB
    11. 10084
    Prev
    Next
    INDEX
    Explanations

    references to a duration of time spanning multiple years

    oai_token-act-pair · gpt-3.5-turbo

    phrases that indicate time durations exceeding a specific threshold

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

    No Comments

    Negative Logits
     Lauder
    -0.77
    ":[
    -0.73
    ãĤ´
    -0.68
    arium
    -0.62
    Ô
    -0.62
    Mini
    -0.62
    DragonMagazine
    -0.62
    illary
    -0.62
    heses
    -0.61
    UES
    -0.61
    POSITIVE LOGITS
    drive
    1.08
    hang
    1.04
    priced
    1.00
    population
    0.96
    whelming
    0.92
    lander
    0.92
    capacity
    0.90
    valued
    0.88
    tones
    0.88
    stated
    0.86
    Activations Density 0.058%

    No Known Activations