© 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. 3-RES-JB
    11. 8728
    Prev
    Next
    INDEX
    Explanations

    instances of the word "Twilight" occurring in the text

    oai_token-act-pair · gpt-3.5-turbo

    references to the character Twilight from the "My Little Pony" franchise

    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
    sembly
    -0.80
    ulative
    -0.79
    Ĩ
    -0.76
    irin
    -0.76
    olulu
    -0.74
    ĪĴ
    -0.74
    alty
    -0.71
    ntil
    -0.71
    atown
    -0.69
    Ģ
    -0.68
    POSITIVE LOGITS
     Sparkle
    1.68
     Twilight
    1.19
     Imperium
    0.93
     Princess
    0.93
     Zone
    0.90
     Struggle
    0.87
     Spike
    0.82
     Celest
    0.77
    ilight
    0.76
    ween
    0.75
    Activations Density 0.012%

    No Known Activations