© 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. 0-RES-JB
    11. 8034
    Prev
    Next
    INDEX
    Explanations

    phrases related to different forms and contexts of light

    oai_token-act-pair · gpt-3.5-turbo

    references to various forms of "light" and its derivatives

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     bogus
    -0.72
    qqa
    -0.64
     coerced
    -0.64
     diplom
    -0.62
     timely
    -0.60
     divor
    -0.59
     Ass
    -0.57
     incorrect
    -0.57
     perm
    -0.57
     transfer
    -0.56
    POSITIVE LOGITS
    light
    4.43
    lights
    3.16
    lighting
    2.75
    Light
    2.12
    lit
    2.04
     light
    1.79
     Light
    1.67
    lite
    1.62
    dark
    1.46
     Lights
    1.27
    Activations Density 0.007%

    No Known Activations