© 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-IT SAEs
    5. Gemma-2B-IT
    7. Residual Stream
    9. 12-RES-JB
    11. 10636
    Prev
    Next
    INDEX
    Explanations

    descriptions of geographic locations and events related to outdoor activities

    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
    +0.15
    0.5%
    906
    +0.13
    0.4%
    1741
    +0.12
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    227
    +0.15
    0.08
    648
    +0.13
    0.06
    283
    +0.12
    0.01
    Negative Logits
     hozz
    -0.65
    ControllerAdvice
    -0.64
     loggedIn
    -0.62
    AccessLevel
    -0.61
     hypotension
    -0.61
     userEmail
    -0.61
     eventName
    -0.59
    šech
    -0.58
     responseData
    -0.57
     spake
    -0.57
    POSITIVE LOGITS
     Chá
    1.19
     Strukt
    1.17
     Anm
    1.15
     Præ
    1.15
     mef
    1.14
     Khart
    1.14
     aen
    1.14
     Juf
    1.13
     hcm
    1.11
     ivi
    1.10
    Activations Density 0.320%

    No Known Activations