© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 11743
    Prev
    Next
    INDEX
    Explanations

    articles and descriptive phrases related to materials and compositions

    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
    156
    +0.32
    1.9%
    21
    +0.21
    1.2%
    478
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    436
    +0.32
    0.01
    507
    +0.21
    0.01
    123
    +0.13
    0.01
    Negative Logits
    mares
    -1.67
    hline
    -1.58
    fall
    -1.53
    ality
    -1.53
    ––––––––
    -1.52
    IDTH
    -1.48
    autres
    -1.47
     Ages
    -1.44
    vier
    -1.43
    áĥĶ
    -1.42
    POSITIVE LOGITS
    ©
    2.50
    »¿
    2.39
    ĸ
    2.37
    ħ
    2.29
    Ļª
    2.28
    ĵ
    2.26
     plurality
    2.18
    Ĵ
    2.14
    Ł
    2.13
    ¢
    2.09
    Activations Density 0.027%

    No Known Activations