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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 12456
    Prev
    Next
    INDEX
    Explanations

    phrases indicating surprise or shock

    oai_token-act-pair · gpt-3.5-turbo

    expressions of shock or disapproval regarding negative events or situations

    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
    •layer 8 - surprise

    No Comments

    Negative Logits
    uay
    -0.84
    arten
    -0.82
    oller
    -0.77
    owned
    -0.71
    check
    -0.71
    isSpecialOrderable
    -0.70
    orno
    -0.69
    abel
    -0.68
    igate
    -0.68
    arate
    -0.68
    POSITIVE LOGITS
     outcome
    1.16
     prospect
    1.03
     portrayal
    1.01
     enorm
    0.99
     plight
    0.98
     lack
    0.98
     notion
    0.97
     idea
    0.97
     situation
    0.97
     possibility
    0.93
    Activations Density 0.269%

    No Known Activations