© 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. 10-RES-JB
    11. 16253
    Prev
    Next
    INDEX
    Explanations

    professionals with specific roles or identities (e.g., author, blogger, coach, philosopher) as well as specific interests or advocacies (e.g., activist, advocate, gamer)

    oai_token-act-pair · gpt-3.5-turbo

    roles and occupations of individuals, particularly those associated with creativity, activism, and professionalism

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     or
    -0.74
    ankind
    -0.67
     AND
    -0.65
     and
    -0.64
    ords
    -0.60
    existing
    -0.60
    erning
    -0.60
     nor
    -0.58
    ModLoader
    -0.58
    pmwiki
    -0.58
    POSITIVE LOGITS
     alike
    1.33
     extraord
    1.30
     respectively
    1.01
     advocate
    0.84
     educator
    0.72
     enthusiast
    0.71
     proponent
    0.68
     specializing
    0.65
     collaborator
    0.64
     hybrid
    0.64
    Activations Density 0.125%

    No Known Activations