© 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 SAEs
    5. Gemma-2B
    7. Residual Stream
    9. 10-RES-JB
    11. 4081
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate significant statistics or amounts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.10.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    chanind/openwebtext-gemma
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    Architecture
    standard
    Context Size
    1,024
    Dataset
    ctigges/openwebtext-gemma-1024-cl
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    50
    +0.27
    1.1%
    1265
    +0.12
    0.5%
    1763
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1516
    +0.27
    0.03
    1983
    +0.12
    0.03
    1416
    +0.10
    0.03
    Negative Logits
    <bos>
    -2.08
    ⓧ
    -0.90
    <?
    -0.65
     meet
    -0.61
     cooperated
    -0.61
     select
    -0.60
     attend
    -0.59
     bring
    -0.57
     met
    -0.57
     announce
    -0.57
    POSITIVE LOGITS
     bordeaux
    1.20
     cabrio
    1.19
     cannes
    1.18
     lele
    1.18
     fatis
    1.14
     Messieurs
    1.13
     milano
    1.13
     affez
    1.12
     ananas
    1.11
     sentra
    1.10
    Activations Density 0.136%

    No Known Activations