Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 8970
    Prev
    Next
    INDEX
    Explanations

    references to a person named Gilligan with different contexts such as a sports commentator, politician, or show participant

    oai_token-act-pair · gpt-3.5-turbo

    mentions of specific individuals, particularly those with the last name "Gilligan."

    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
    Sync
    -0.81
     iso
    -0.73
    sync
    -0.72
     Sync
    -0.72
    seek
    -0.71
    ighters
    -0.71
     sync
    -0.69
     flares
    -0.69
    weekly
    -0.68
    Beta
    -0.67
    POSITIVE LOGITS
     Gill
    3.87
     Rudd
    1.58
     Gerry
    1.19
     Kirst
    1.16
     Gins
    1.15
     Gillespie
    1.14
     Gina
    1.12
     Dill
    1.12
     Jill
    1.11
     Garg
    1.11
    Activations Density 0.022%

    No Known Activations