© 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. 19781
    Prev
    Next
    INDEX
    Explanations

    terms related to legal judgments or rulings

    oai_token-act-pair · gpt-3.5-turbo

    terms associated with public health and safety

    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
     crappy
    -0.48
     Slayer
    -0.47
     haha
    -0.47
     shitty
    -0.47
     cheesy
    -0.45
     dude
    -0.44
    Redditor
    -0.44
     fucking
    -0.43
     crap
    -0.43
     horny
    -0.43
    POSITIVE LOGITS
     meanwhile
    0.51
     briefed
    0.50
    classified
    0.48
     embold
    0.48
     Tillerson
    0.47
    zbek
    0.47
     redacted
    0.47
    ccording
    0.46
     cited
    0.46
     Polit
    0.46
    Activations Density 3.468%

    No Known Activations