© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 9237
    Prev
    Next
    INDEX
    Explanations

    surface-related concepts or descriptions

    oai_token-act-pair · gpt-3.5-turbo

    references to the surface of various entities or phenomena

    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
    avering
    -0.80
    agne
    -0.76
    agna
    -0.75
    ér
    -0.73
    eros
    -0.72
    rams
    -0.72
    anders
    -0.71
    erie
    -0.70
    allo
    -0.70
    orum
    -0.69
    POSITIVE LOGITS
    tenance
    0.94
    FACE
    0.83
     surface
    0.75
    cloth
    0.73
     surfaces
    0.71
    faces
    0.69
     layer
    0.69
     combatants
    0.69
     Surface
    0.68
     cleaners
    0.68
    Activations Density 0.025%

    No Known Activations