© 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. 1903
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal matters and strong opinions expressed in a forceful or alarming way

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    The neuron fires on third-person singular present-tense verbs (the “-s” verbs like “touches,” “dies,” “spends,” “happens,” “is,” etc.).

    oai_token-act-pair · o4-miniTriggered by @phamvanluong2003

    The neuron detects the present‑tense copula "is" (third‑person singular form of the verb "to be").

    oai_token-act-pair · gpt-5-miniTriggered by @phamvanluong2003
    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
    266
    +0.08
    0.2%
    297
    +0.08
    0.2%
    1241
    +0.08
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    713
    +0.08
    0.04
    256
    +0.08
    0.03
    963
    +0.08
    0.03
    Negative Logits
     Sumatera
    -0.60
     Jambi
    -0.57
     kafe
    -0.56
     Tangerang
    -0.54
     Lampung
    -0.52
     jati
    -0.51
     induk
    -0.51
     Kalimantan
    -0.51
     Banten
    -0.50
     karton
    -0.49
    POSITIVE LOGITS
     lamborghini
    0.81
     cushi
    0.77
     shewn
    0.72
     hairc
    0.70
     swarovski
    0.70
     tupperware
    0.69
     feen
    0.69
     chrysler
    0.69
     vhs
    0.69
     scrat
    0.68
    Activations Density 0.368%

    No Known Activations