© 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. 11678
    Prev
    Next
    INDEX
    Explanations

    instances of the word "ent" across various contexts

    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
    271
    +0.16
    0.9%
    21
    +0.12
    0.7%
    199
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    236
    +0.16
    0.02
    5
    +0.12
    0.01
    199
    +0.11
    0.02
    Negative Logits
     sense
    -1.73
    ths
    -1.70
    etin
    -1.53
    fully
    -1.48
     same
    -1.45
     presence
    -1.44
    backs
    -1.40
    fulness
    -1.38
    th
    -1.36
    mente
    -1.35
    POSITIVE LOGITS
    Ŀ
    2.66
    ŀ
    2.44
    ¶
    2.38
    IJ
    2.33
    ¯
    2.24
    Ļ
    2.21
    ı
    2.21
    Ń
    2.20
    º
    2.20
    ¨
    2.19
    Activations Density 0.067%

    No Known Activations