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

    Neuronpedia

    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. 3933
    Prev
    Next
    INDEX
    Explanations

    mentions of items and their characteristics like toys, metal, plastic, and riffs in a descriptive narrative

    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
    198
    +0.14
    0.4%
    674
    +0.11
    0.3%
    1967
    +0.10
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.14
    0.06
    683
    +0.11
    0.06
    1373
    +0.10
    0.01
    Negative Logits
     apprehen
    -0.63
     Wtf
    -0.59
     unspeak
    -0.58
     ItemType
    -0.57
     somethin
    -0.56
     logitech
    -0.56
     shenan
    -0.56
     fucker
    -0.54
     Lmao
    -0.53
     tille
    -0.52
    POSITIVE LOGITS
     habang
    0.64
     kayo
    0.64
     kapag
    0.63
     kasama
    0.63
     acquistare
    0.62
     ngayon
    0.62
     parteci
    0.61
     offrire
    0.61
    borderTop
    0.61
     maraming
    0.61
    Activations Density 0.411%

    No Known Activations