© 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. 9535
    Prev
    Next
    INDEX
    Explanations

    statements describing "seen" or related terms like "saw" or "heard"

    oai_token-act-pair · gpt-3.5-turbo

    expressions indicating unique or unprecedented experiences

    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
     portfolios
    -0.81
     antioxid
    -0.67
    anwhile
    -0.63
    Alert
    -0.61
     Supplemental
    -0.61
     Actions
    -0.60
    Solution
    -0.60
     ASAP
    -0.59
     Vulkan
    -0.58
     Luther
    -0.56
    POSITIVE LOGITS
     slightest
    0.88
     EVER
    0.85
     dreamed
    0.82
     nor
    0.81
     anywhere
    0.79
    ukong
    0.78
     dime
    0.77
    ratch
    0.76
     whatsoever
    0.76
    thood
    0.73
    Activations Density 0.226%

    No Known Activations