© 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. 12725
    Prev
    Next
    INDEX
    Explanations

    numerical values associated with errors or thresholds

    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
    78
    +0.24
    1.5%
    9
    +0.18
    1.2%
    348
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    9
    +0.24
    0.08
    78
    +0.18
    0.02
    273
    +0.13
    0.05
    Negative Logits
    icro
    -1.54
    ense
    -1.48
     \[[
    -1.35
    âĪ¶
    -1.31
     vectors
    -1.31
    ching
    -1.22
    imately
    -1.20
    scar
    -1.20
    pes
    -1.18
    CMS
    -1.18
    POSITIVE LOGITS
    Īĺ
    3.78
    Ľ
    3.68
    ŀ
    3.63
    į
    3.56
    ©
    3.38
    ³
    3.37
    Ĥ
    3.35
    ł
    3.34
    Ĭ
    3.30
    ĺ
    3.29
    Activations Density 0.431%

    No Known Activations