© 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. 13600
    Prev
    Next
    INDEX
    Explanations

    references to string variables and associated operations in code

    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
    25
    +0.20
    1.1%
    220
    +0.13
    0.7%
    4
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    220
    +0.20
    0.03
    254
    +0.13
    0.01
    481
    +0.13
    0.05
    Negative Logits
    zilla
    -1.44
    golang
    -1.43
     <-
    -1.30
    tox
    -1.29
     [@
    -1.28
    craft
    -1.27
    ulent
    -1.25
    $$
    -1.25
    ulence
    -1.24
    HAVE
    -1.24
    POSITIVE LOGITS
     second
    4.79
     third
    4.34
     fourth
    4.09
     Second
    3.97
    Second
    3.85
    second
    3.74
     fifth
    3.73
     sixth
    3.71
     Secondly
    3.51
    third
    3.45
    Activations Density 1.026%

    No Known Activations