© 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. 11-RES-JB
    11. 5489
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical actions and bodily sensations, such as movements and feelings

    oai_token-act-pair · gpt-3.5-turbo

    instances of intensity or emphasis in actions and descriptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     commence
    -0.82
     realise
    -0.77
     are
    -0.76
     aren
    -0.75
     clarify
    -0.73
     realize
    -0.70
    igate
    -0.70
     ARE
    -0.70
     arrive
    -0.68
     determine
    -0.68
    POSITIVE LOGITS
     drew
    1.10
     wore
    1.05
     lacked
    1.05
     ate
    1.03
     sank
    1.02
     went
    1.02
     breathed
    1.02
     drank
    1.02
     tasted
    1.01
     smelled
    1.01
    Activations Density 0.662%

    No Known Activations