© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Initial Gemma-2B SAEs
    5. Gemma-2B
    7. Residual Stream
    9. 0-RES-JB
    11. 1269
    Prev
    Next
    INDEX
    Explanations

    phrases associated with humor or joking

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    1178
    +0.10
    0.3%
    913
    +0.09
    0.2%
    486
    +0.08
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1363
    +0.10
    0.04
    690
    +0.09
    0.05
    283
    +0.08
    0.01
    Negative Logits
     intptr
    -0.72
     invokingState
    -0.57
    SourceChecksum
    -0.56
    HasBeenSet
    -0.55
    webElementXpaths
    -0.54
     oprot
    -0.54
     onOptions
    -0.53
    ("")]
    -0.52
    sidemargin
    -0.52
    tre
    -0.52
    POSITIVE LOGITS
     Peters
    1.23
    Peters
    1.14
     kidding
    1.14
     vns
    0.97
     lamborghini
    0.96
     Augu
    0.95
     PETERS
    0.94
     prae
    0.91
     Amerik
    0.90
     skechers
    0.89
    Activations Density 0.325%

    No Known Activations