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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 4-RES-JB
    11. 13714
    Prev
    Next
    INDEX
    Explanations

    phrases related to taking action against something

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to enforcement actions or measures against something

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

    No Comments

    Negative Logits
    ãĤ¤ãĥĪ
    -0.70
     sake
    -0.68
    æµ
    -0.65
    ubes
    -0.62
     Norn
    -0.61
     Upton
    -0.61
     amen
    -0.61
     embraces
    -0.60
    enburg
    -0.60
     pronoun
    -0.60
    POSITIVE LOGITS
     wiret
    0.83
    orously
    0.80
    riminal
    0.79
     prosecuted
    0.72
    ascus
    0.71
    Operation
    0.71
    criminal
    0.68
     eaves
    0.68
     Illegal
    0.67
    afia
    0.67
    Activations Density 0.051%

    No Known Activations