© 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. 16739
    Prev
    Next
    INDEX
    Explanations

    website URLs

    oai_token-act-pair · gpt-3.5-turbo

    URLs and web addresses 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
    zona
    -0.78
     Wonderland
    -0.77
     Phi
    -0.71
     Debor
    -0.70
     Glac
    -0.67
     Excellence
    -0.66
     Haas
    -0.63
     Dahl
    -0.63
     caucuses
    -0.61
    quartered
    -0.61
    POSITIVE LOGITS
    youtube
    1.43
    facebook
    1.32
    reddit
    1.15
    ebin
    0.96
    scrib
    0.92
    google
    0.92
    kick
    0.89
    cms
    0.88
    github
    0.86
    etsy
    0.86
    Activations Density 0.035%

    No Known Activations