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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 29307
    Prev
    Next
    INDEX
    Explanations

    terms related to educational and training structures in manufacturing

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    198
    +0.15
    0.9%
    34
    +0.14
    0.8%
    312
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.15
    0.12
    312
    +0.14
    0.13
    418
    +0.13
    -0.04
    Negative Logits
    ĻĤ
    -3.22
    »¿
    -2.98
    ĨĴ
    -2.96
    Īĺ
    -2.60
    ·¸
    -2.55
    ĸ´
    -2.35
    Ī
    -2.33
    ĺ
    -2.12
    ĭ
    -2.11
    ľĵ
    -2.11
    POSITIVE LOGITS
    âĢIJ
    1.70
    metrics
    1.65
     cert
    1.64
     edition
    1.58
     formul
    1.54
     and
    1.52
    -
    1.51
    amental
    1.50
     domain
    1.47
    istical
    1.47
    Activations Density 4.016%

    No Known Activations