© 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. 11-RES-JB
    11. 5306
    Prev
    Next
    INDEX
    Explanations

    terms related to detention or holding someone in custody

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "Det" or variations thereof that indicate a focus on detention

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    manship
    -0.82
     bold
    -0.64
    #$
    -0.60
     grave
    -0.57
    xual
    -0.56
    ãģĦ
    -0.56
     Tobacco
    -0.56
     Zamb
    -0.56
    BSD
    -0.55
     cosmos
    -0.55
    POSITIVE LOGITS
    achable
    1.46
    roit
    1.41
    ailed
    1.22
    ention
    1.17
    ective
    1.16
    achment
    1.15
    ainer
    1.15
    rans
    1.14
    ached
    1.14
    ector
    1.14
    Activations Density 0.033%

    No Known Activations