© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Initial Gemma-2B-IT SAEs
    5. Gemma-2B-IT
    7. Residual Stream
    9. 12-RES-JB
    11. 5549
    Prev
    Next
    INDEX
    Explanations

    technical instructions for software installation or coding tutorials

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    the neuron detects instructional/explanatory/documentation language (words that signal tutorials, introductions, explanations, API/method references, or "read/see here" guidance).

    oai_token-act-pair · gpt-5-miniTriggered by @phamvanluong2003
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Hook Layer
    12
    Architecture
    standard
    Context Size
    1,024
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    876
    +0.23
    0.7%
    872
    +0.20
    0.6%
    1699
    +0.09
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    872
    +0.23
    0.08
    876
    +0.20
    -0.02
    350
    +0.09
    0.03
    Negative Logits
     konserv
    -0.82
     kriminal
    -0.81
     Demokrat
    -0.79
     foton
    -0.79
     indik
    -0.78
     republi
    -0.76
     kompati
    -0.74
     panik
    -0.73
     kontinu
    -0.72
     demokrat
    -0.72
    POSITIVE LOGITS
     ennemi
    0.74
     automne
    0.69
     écout
    0.67
     malheur
    0.65
     pylab
    0.64
     pymysql
    0.64
     Explained
    0.63
     joyeux
    0.63
     tupperware
    0.62
     mépris
    0.62
    Activations Density 0.551%

    No Known Activations