© 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. 2840
    Prev
    Next
    INDEX
    Explanations

    verbs ending in 'ing' or adjectives ending in 'ack', 'aps', 'ep', 'ed', 'ing', or 'unk'

    oai_token-act-pair · gpt-3.5-turbo

    colloquial or informal expressions of negative interactions or sentiments

    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
     Colo
    -0.66
     cryst
    -0.62
     SLI
    -0.61
    OTOS
    -0.61
    IJ
    -0.59
    Pa
    -0.58
    izoph
    -0.58
    SW
    -0.58
    URA
    -0.57
     CRE
    -0.57
    POSITIVE LOGITS
    creen
    0.93
    iness
    0.93
    weed
    0.91
    ucker
    0.90
    buck
    0.90
    ily
    0.89
    ishly
    0.88
    manship
    0.87
     havoc
    0.84
    stakes
    0.84
    Activations Density 0.497%

    No Known Activations