© 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. 8-RES-JB
    11. 6040
    Prev
    Next
    INDEX
    Explanations

    words related to conventions or established norms

    oai_token-act-pair · gpt-3.5-turbo

    instances of the prefix "Con" in words, typically indicating concepts related to conditions or agreements

    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
     Wem
    -0.68
    gypt
    -0.63
     wip
    -0.63
    assetsadobe
    -0.61
    OHN
    -0.61
     heights
    -0.60
     Rover
    -0.60
     Bulg
    -0.59
     checkpoints
    -0.58
    aceutical
    -0.57
    POSITIVE LOGITS
    ventional
    1.31
    crete
    1.29
    stant
    1.28
    venient
    1.28
    cern
    1.26
    ventions
    1.25
    structed
    1.25
    cerned
    1.25
    clusions
    1.24
    secut
    1.22
    Activations Density 0.034%

    No Known Activations