© 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. 0-RES-JB
    11. 22394
    Prev
    Next
    INDEX
    Explanations

    instances of the word "mocked"

    oai_token-act-pair · gpt-3.5-turbo

    references to mocking or ridicule

    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
     UNCLASSIFIED
    -0.78
    Benz
    -0.77
     AAA
    -0.76
     Concent
    -0.76
    HQ
    -0.76
     conserv
    -0.76
    impact
    -0.73
    clinton
    -0.72
     nexus
    -0.70
     determin
    -0.69
    POSITIVE LOGITS
     mocked
    2.95
     mocking
    2.69
     mockery
    2.27
     teasing
    2.14
     ridicule
    2.10
     teased
    2.10
     ridiculed
    2.01
     mock
    1.81
     ridic
    1.74
     joking
    1.70
    Activations Density 0.048%

    No Known Activations