© 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. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 1-RES-SM
    11. 24499
    Prev
    Next
    INDEX
    Explanations

    frequent usage of the word "the" in various contexts

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    40
    +0.14
    0.8%
    507
    +0.11
    0.6%
    161
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    10
    +0.14
    0.01
    98
    +0.11
    0.00
    23
    +0.10
    0.01
    Negative Logits
    ķ
    -1.97
    ĵ
    -1.88
    Ģ
    -1.87
    Ĭ
    -1.85
    ĥ
    -1.82
    ¤
    -1.76
    ĻĤ
    -1.76
    »¿
    -1.76
    ¿
    -1.70
    ¾
    -1.68
    POSITIVE LOGITS
    ulsions
    1.53
     Appeal
    1.46
    )",
    1.43
    apore
    1.41
    infection
    1.38
     Virus
    1.36
     Response
    1.35
    reaction
    1.33
    router
    1.33
    brit
    1.29
    Activations Density 0.005%

    No Known Activations