© 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
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 4-RES-JB
    6. 22412
    Prev
    Next
    INDEX
    Explanations

    phrases related to sweeping actions or activities

    oai_token-act-pair · gpt-3.5-turbo

    forms of the word "sweep" and related concepts of sweeping actions

    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
     willpower
    -0.68
     Palestin
    -0.64
    cil
    -0.64
     restraint
    -0.63
    ña
    -0.63
     gratification
    -0.62
    ourke
    -0.61
     kneeling
    -0.61
    acters
    -0.60
     Founding
    -0.59
    POSITIVE LOGITS
    stakes
    1.75
    restling
    0.96
     sweep
    0.86
    opa
    0.83
    brush
    0.82
    shots
    0.80
    itzer
    0.79
     sweeping
    0.78
    hurst
    0.78
    ãĥĥãĥĪ
    0.77
    Activations Density 0.013%

    No Known Activations