© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 11527
    Prev
    Next
    INDEX
    Explanations

    references to governmental entities and their various jurisdictions

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    309
    +0.14
    0.8%
    246
    +0.14
    0.8%
    119
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    309
    +0.14
    0.04
    119
    +0.14
    0.03
    168
    +0.13
    0.04
    Negative Logits
    Ĭ
    -3.40
    į
    -3.37
    ·¸
    -3.35
    ½
    -3.23
    ĭ
    -3.14
    ij
    -3.11
    ĥ½
    -3.06
    -3.04
    ↵
    -3.04
    č↵
    -3.04
    POSITIVE LOGITS
    wide
    2.57
     legislature
    2.56
     Department
    2.38
    craft
    2.31
    houses
    2.30
    house
    2.26
    hood
    2.22
     legisl
    2.18
    building
    2.14
     Legislature
    2.14
    Activations Density 0.148%

    No Known Activations