© 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. 22455
    Prev
    Next
    INDEX
    Explanations

    words related to titles or names combined with descriptions like "Spear of Longinus" or "Prince of Dorne."

    oai_token-act-pair · gpt-3.5-turbo

    Titles and names of media, particularly related to dramas and fantasy genres

    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
     reperto
    -0.69
     bandwagon
    -0.65
     weekday
    -0.65
     promoter
    -0.64
     commercially
    -0.63
     casualty
    -0.62
     cutoff
    -0.62
     repertoire
    -0.62
     trough
    -0.62
    arnaev
    -0.61
    POSITIVE LOGITS
    areth
    0.90
    hent
    0.85
    nesia
    0.83
     fame
    0.82
    minster
    0.78
    thood
    0.77
    oneliness
    0.74
    arth
    0.74
    mort
    0.73
    ropy
    0.72
    Activations Density 0.317%

    No Known Activations