© 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. 8-RES-JB
    11. 8061
    Prev
    Next
    INDEX
    Explanations

    repetitions of actions or events

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "again."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ocene
    -0.65
     Nicarag
    -0.64
    regate
    -0.62
     Venezuelan
    -0.62
     Payton
    -0.61
    kees
    -0.59
    opol
    -0.58
     tents
    -0.58
    EMP
    -0.58
     Egyptians
    -0.58
    POSITIVE LOGITS
    nir
    0.81
     actionGroup
    0.75
    abl
    0.74
    ilver
    0.71
    nces
    0.67
     conclud
    0.67
    EngineDebug
    0.66
     tremend
    0.66
    hots
    0.65
    orks
    0.65
    Activations Density 0.031%

    No Known Activations