Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Initial Gemma-2B-IT SAEs
    3. Gemma-2B-IT
    4. Residual Stream
    5. 12-RES-JB
    6. 8366
    Prev
    Next
    INDEX
    Explanations

    contractions of verbs with negative indicators

    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
    381
    +0.12
    0.4%
    1557
    +0.10
    0.3%
    1077
    +0.10
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1557
    +0.12
    0.03
    1604
    +0.10
    0.02
    689
    +0.10
    0.02
    Negative Logits
     makro
    -1.12
     akut
    -1.10
     solidar
    -1.04
     ideolog
    -1.01
     kac
    -1.00
     alkoh
    -0.97
     antik
    -0.97
     buk
    -0.97
     teras
    -0.96
     optik
    -0.96
    POSITIVE LOGITS
     unspeak
    1.16
     tolerably
    1.09
     gaily
    1.08
     unwarran
    1.04
     apprehen
    1.03
     intersper
    1.02
     indestru
    1.01
     disagre
    1.01
     shenan
    0.99
     Shakspeare
    0.98
    Activations Density 0.175%

    No Known Activations