© 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. 11256
    Prev
    Next
    INDEX
    Explanations

    accusatory language and intense emotional statements related to personal relationships and interactions

    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
    185
    +0.09
    0.2%
    270
    +0.08
    0.2%
    604
    +0.07
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    915
    +0.09
    0.04
    972
    +0.08
    0.03
    1951
    +0.07
    0.04
    Negative Logits
     accla
    -1.18
     effe
    -1.13
     emphat
    -1.10
     increa
    -1.09
     coö
    -1.08
     encomp
    -1.06
     fta
    -1.05
     inev
    -1.05
     ?...
    -1.04
     invin
    -1.04
    POSITIVE LOGITS
    NEVER
    0.82
     NEVER
    0.81
    Never
    0.79
     never
    0.78
     Never
    0.76
    never
    0.75
    posterous
    0.67
     wouldn
    0.67
     unthinkable
    0.61
     WOULD
    0.57
    Activations Density 0.471%

    No Known Activations