© 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. 8-RES-JB
    11. 16908
    Prev
    Next
    INDEX
    Explanations

    mathematical and programming-related symbols and terms

    oai_token-act-pair · gpt-3.5-turbo

    repeated use of specific symbols or characters in the text

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     warr
    -0.87
     general
    -0.73
     ende
    -0.70
     nude
    -0.69
     elig
    -0.68
     unemployed
    -0.68
     myster
    -0.67
     veget
    -0.66
     independ
    -0.66
     facult
    -0.65
    POSITIVE LOGITS
     Whereas
    0.96
     Unfortunately
    0.94
    Though
    0.90
     Sadly
    0.89
     Thankfully
    0.88
     Certainly
    0.88
    However
    0.88
    Ironically
    0.88
     Granted
    0.87
     Because
    0.86
    Activations Density 0.207%

    No Known Activations