© 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. 22623
    Prev
    Next
    INDEX
    Explanations

    the word "trick" with high activation

    oai_token-act-pair · gpt-3.5-turbo

    references to tricks or deceptive strategies

    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
    angan
    -0.79
    ews
    -0.70
    omes
    -0.67
    Ê
    -0.66
     Plaint
    -0.66
    EO
    -0.64
    omic
    -0.63
     Prior
    -0.63
    Dur
    -0.62
     Afric
    -0.62
    POSITIVE LOGITS
     trick
    3.92
     tricks
    2.49
     Trick
    2.23
     prank
    1.49
     fool
    1.38
     gimmick
    1.32
     tricked
    1.32
     stunt
    1.24
     tactic
    1.19
     tricky
    1.18
    Activations Density 0.022%

    No Known Activations