Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Initial Gemma-2B-IT SAEs
    3. Gemma-2B-IT
    4. Residual Stream
    5. 12-RES-JB
    6. 10011
    Prev
    Next
    INDEX
    Explanations

    processes and methods related to material production, such as grinding, dissolving, and storing

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    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
    1013
    +0.07
    0.2%
    184
    +0.07
    0.2%
    1038
    +0.07
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    181
    +0.07
    0.03
    1038
    +0.07
    0.03
    1014
    +0.07
    0.04
    Negative Logits
     kram
    -1.13
     lemp
    -1.10
     alkoh
    -1.10
     solidar
    -1.09
     uhr
    -1.05
     utop
    -1.00
     menta
    -0.98
     maksi
    -0.98
     saar
    -0.97
     Ukraina
    -0.97
    POSITIVE LOGITS
     subsequently
    0.81
     then
    0.80
     następnie
    0.69
    then
    0.68
     later
    0.67
     затем
    0.64
     vervolgens
    0.63
     THEN
    0.63
     luego
    0.62
     wiederum
    0.59
    Activations Density 0.334%

    No Known Activations