© 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. 4743
    Prev
    Next
    INDEX
    Explanations

    words related to inefficiency or lack of usefulness

    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
    1363
    +0.09
    0.2%
    2030
    +0.08
    0.2%
    1013
    +0.08
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    2030
    +0.09
    0.05
    1615
    +0.08
    0.06
    1315
    +0.08
    0.04
    Negative Logits
     emphat
    -0.72
     wattpad
    -0.66
     pessi
    -0.63
     solidar
    -0.62
     suspic
    -0.61
     fuf
    -0.59
     disagre
    -0.59
     proprement
    -0.59
     intrigu
    -0.59
     chiaramente
    -0.58
    POSITIVE LOGITS
     unless
    0.83
    unless
    0.74
     except
    0.72
    except
    0.66
    Unless
    0.58
     Unless
    0.55
     compared
    0.55
    Except
    0.51
     useless
    0.51
    hodné
    0.48
    Activations Density 0.313%

    No Known Activations