© 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. 14663
    Prev
    Next
    INDEX
    Explanations

    instances of the word 'and' used to connect ideas or phrases

    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
    +0.20
    1.1%
    421
    +0.15
    0.8%
    348
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    289
    +0.20
    0.03
    237
    +0.15
    0.04
    495
    +0.13
    0.04
    Negative Logits
    maven
    -1.54
    µ
    -1.53
    dorff
    -1.44
    ¢
    -1.44
    £
    -1.42
    (@"
    -1.41
     hierarchy
    -1.35
    »¿
    -1.34
    ("
    -1.33
    artifactId
    -1.33
    POSITIVE LOGITS
    rogens
    3.03
    rog
    2.24
    rew
    2.06
    rogen
    2.06
    yz
    1.96
    rost
    1.94
    rogenic
    1.91
     respectively
    1.81
     behold
    1.80
    onas
    1.79
    Activations Density 0.229%

    No Known Activations