© 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. 3-RES-JB
    11. 12442
    Prev
    Next
    INDEX
    Explanations

    words related to a significant failure or breakdown

    oai_token-act-pair · gpt-3.5-turbo

    references to various types of collapses, particularly in economic or structural contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    qua
    -0.75
    friends
    -0.71
    ĭ
    -0.71
    lean
    -0.70
    ña
    -0.68
    Choice
    -0.67
    cil
    -0.67
    tein
    -0.65
    tes
    -0.65
    ju
    -0.65
    POSITIVE LOGITS
     collapse
    0.93
     inertia
    0.84
     downfall
    0.80
    ulence
    0.80
     collapses
    0.78
     contag
    0.78
    nyder
    0.78
    ibility
    0.77
    opian
    0.76
     tremend
    0.76
    Activations Density 0.018%

    No Known Activations