© 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 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. 8-RES-JB
    11. 10412
    Prev
    Next
    INDEX
    Explanations

    the word 'witness' along with the context in which it appears

    oai_token-act-pair · gpt-3.5-turbo

    references to witnesses in legal or investigative contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ategory
    -0.73
     McMaster
    -0.73
     Pis
    -0.71
    scill
    -0.68
    oku
    -0.67
    ofi
    -0.67
    matically
    -0.66
     Chop
    -0.66
    angu
    -0.66
    Saharan
    -0.65
    POSITIVE LOGITS
     testimonies
    1.03
    ewitness
    0.99
     testimony
    0.96
     witness
    0.95
     testify
    0.94
    Witness
    0.90
     Witnesses
    0.88
     witnesses
    0.88
     tampering
    0.86
     testifying
    0.85
    Activations Density 0.026%

    No Known Activations