© 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. 3-RES-JB
    11. 6531
    Prev
    Next
    INDEX
    Explanations

    phrases related to personal growth and upbringing

    oai_token-act-pair · gpt-3.5-turbo

    the phrase "growing up."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     mined
    -0.69
    eport
    -0.68
    mination
    -0.66
    mir
    -0.64
    mouth
    -0.63
    mining
    -0.61
    uminati
    -0.61
     disclosure
    -0.60
    phis
    -0.59
     mitigation
    -0.59
    POSITIVE LOGITS
    landish
    0.84
    bryce
    0.82
    MpServer
    0.77
    adesh
    0.74
    skirts
    0.74
    ãĤ¼ãĤ¦ãĤ¹
    0.73
     redes
    0.72
    ris
    0.71
    stairs
    0.68
    ISTER
    0.68
    Activations Density 0.041%

    No Known Activations