© 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. 0-RES-JB
    11. 15090
    Prev
    Next
    INDEX
    Explanations

    phrases related to being physically impacted or struck by an external force

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "knocked."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Fein
    -0.78
     Eth
    -0.68
     Chart
    -0.65
     Ceres
    -0.63
     Advent
    -0.62
    union
    -0.62
     Accord
    -0.61
     Exodus
    -0.61
    Abyss
    -0.61
     Argentina
    -0.59
    POSITIVE LOGITS
     knocked
    3.33
     knocks
    2.08
     knocking
    2.02
     knock
    1.98
     toppled
    1.58
     downed
    1.56
     punched
    1.54
     Knock
    1.45
     pounded
    1.39
     demolished
    1.38
    Activations Density 0.018%

    No Known Activations