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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 2366
    Prev
    Next
    INDEX
    Explanations

    phrases related to close-up views or observations

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "up" and related forms

    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
    owan
    -0.74
     Sturgeon
    -0.70
    conn
    -0.62
    "]=>
    -0.61
    hare
    -0.60
     Yards
    -0.60
    0000000000000000
    -0.59
    xes
    -0.59
     Okin
    -0.58
    [/
    -0.58
    POSITIVE LOGITS
    shot
    0.75
    legged
    0.74
    shots
    0.72
    church
    0.67
    empt
    0.66
    hedon
    0.64
     redes
    0.64
    lier
    0.64
    skirts
    0.64
    LOAD
    0.61
    Activations Density 0.067%

    No Known Activations