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. 14010
    Prev
    Next
    INDEX
    Explanations

    unintended thoughts, memories, and moments that keep resurfacing in the mind

    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
    964
    +0.12
    0.4%
    1013
    +0.10
    0.3%
    674
    +0.10
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1415
    +0.12
    0.03
    194
    +0.10
    0.03
    1267
    +0.10
    0.05
    Negative Logits
     swarovski
    -1.05
     hairc
    -1.03
     softshell
    -0.91
     ecru
    -0.90
    Xoxo
    -0.90
     unlaw
    -0.89
     tupperware
    -0.83
     Confu
    -0.83
     embodi
    -0.82
     velour
    -0.81
    POSITIVE LOGITS
    <bos>
    0.84
     myself
    0.70
     reminded
    0.68
     spontan
    0.67
     reger
    0.63
     besta
    0.62
     realis
    0.62
     handels
    0.61
     intrig
    0.59
     wondering
    0.58
    Activations Density 0.827%

    No Known Activations