© 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. 9067
    Prev
    Next
    INDEX
    Explanations

    phrases that involve referencing and repetition

    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
    248
    +0.13
    0.7%
    371
    +0.13
    0.7%
    477
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    371
    +0.13
    0.05
    248
    +0.13
    0.04
    395
    +0.13
    0.04
    Negative Logits
    shire
    -1.98
    backs
    -1.79
    tre
    -1.66
    s
    -1.58
    worth
    -1.56
     bear
    -1.53
    meg
    -1.52
     affairs
    -1.51
     themselves
    -1.50
    doms
    -1.44
    POSITIVE LOGITS
    ·¸
    3.33
    ³
    3.15
    »
    3.13
    µ
    3.07
    ¬
    3.07
    ·
    3.04
    ´
    2.93
    ¼
    2.86
    ¹
    2.84
    ¯
    2.84
    Activations Density 0.081%

    No Known Activations