© 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. 7-RES-JB
    11. 11112
    Prev
    Next
    INDEX
    Explanations

    proper nouns related to individuals involved in campaigns or organizations

    oai_token-act-pair · gpt-3.5-turbo

    keywords related to concepts of hope and truth

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    aton
    -0.75
    yright
    -0.74
    ounter
    -0.73
    esters
    -0.70
    astical
    -0.69
    ĻĤ
    -0.68
    achy
    -0.65
     discoveries
    -0.65
    atable
    -0.64
    AMES
    -0.64
    POSITIVE LOGITS
    ful
    1.26
    Works
    0.95
    bringer
    0.92
     Matters
    0.92
     Prize
    0.90
     Knot
    0.88
     Bastard
    0.87
    fully
    0.86
     Pact
    0.86
     Belt
    0.85
    Activations Density 0.146%

    No Known Activations