© 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. 10151
    Prev
    Next
    INDEX
    Explanations

    terms related to biochemistry and cellular processes

    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
    158
    +0.13
    0.7%
    397
    +0.12
    0.7%
    498
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    7
    +0.13
    0.02
    354
    +0.12
    0.04
    398
    +0.12
    0.01
    Negative Logits
    _."
    -2.12
    .").
    -2.03
    rangle
    -2.00
    .","
    -1.96
    )."
    -1.95
    .”).
    -1.92
    )?
    -1.89
    ``
    -1.78
    .");
    -1.78
     yet
    -1.77
    POSITIVE LOGITS
     consisted
    3.25
     was
    3.19
     were
    2.98
     ranged
    2.90
     remained
    2.73
     showed
    2.58
     lasted
    2.48
     differed
    2.39
     appeared
    2.36
     amounted
    2.33
    Activations Density 0.405%

    No Known Activations