© 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. 13222
    Prev
    Next
    INDEX
    Explanations

    expressions indicating offering or promotion

    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
    443
    +0.12
    0.7%
    308
    +0.12
    0.7%
    263
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    443
    +0.12
    0.04
    115
    +0.12
    0.03
    308
    +0.11
    0.03
    Negative Logits
    ogens
    -1.49
    IZE
    -1.46
    ĩ
    -1.44
     morning
    -1.43
    ¢
    -1.43
    Ī
    -1.36
     commune
    -1.36
    versus
    -1.30
    thur
    -1.28
     hereby
    -1.28
    POSITIVE LOGITS
    ences
    2.77
    loading
    2.69
    loaded
    2.68
    ensively
    2.65
    shoot
    2.41
    ensive
    2.38
    enders
    2.29
    ices
    2.17
    cuts
    2.10
    screen
    2.04
    Activations Density 0.209%

    No Known Activations