© 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. 4-RES-SM
    11. 9636
    Prev
    Next
    INDEX
    Explanations

    code inclusion directives from programming languages

    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
    397
    +0.14
    0.8%
    271
    +0.12
    0.7%
    9
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    273
    +0.14
    0.04
    144
    +0.12
    0.04
    456
    +0.12
    0.06
    Negative Logits
    oretic
    -1.85
    akers
    -1.51
    irubin
    -1.50
    rav
    -1.49
     certainty
    -1.49
    bia
    -1.45
    herty
    -1.38
    ento
    -1.38
    acin
    -1.36
    otin
    -1.34
    POSITIVE LOGITS
    ŀ
    4.74
    Ī
    4.34
    ¸
    4.29
    ¤
    4.15
    Ĵ
    4.13
    ¿
    4.11
    Ł
    4.11
    ²
    4.10
    ¢
    4.05
    »
    4.05
    Activations Density 0.269%

    No Known Activations