© 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. 6-RES-JB
    11. 3083
    Prev
    Next
    INDEX
    Explanations

    references to scientific studies, grants, and research projects

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.6.hook_resid_post_16384_anthropic_fast_lr
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.6.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    6
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    50
    +0.12
    0.7%
    1339
    +0.11
    0.6%
    1978
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1978
    +0.12
    0.09
    1837
    +0.11
    0.06
    382
    +0.10
    0.06
    Negative Logits
    <bos>
    -2.46
     intersper
    -1.51
     reconno
    -1.21
     shenan
    -1.13
     indestru
    -1.08
    ⓧ
    -1.08
     encomp
    -1.08
     overcrow
    -1.05
     endow
    -1.04
     vainly
    -1.03
    POSITIVE LOGITS
     mikrofon
    0.58
     kristal
    0.58
     protokol
    0.56
     kredi
    0.56
    8
    0.52
     duk
    0.51
     materie
    0.50
     terape
    0.50
     pelu
    0.50
     terapi
    0.49
    Activations Density 0.293%

    No Known Activations