© 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 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. 12459
    Prev
    Next
    INDEX
    Explanations

    occurrences of legal terminology and references

    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
    307
    +0.44
    2.9%
    156
    +0.28
    1.9%
    306
    +0.25
    1.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    307
    +0.44
    0.15
    306
    +0.28
    0.05
    447
    +0.25
    0.09
    Negative Logits
    -2.25
    č↵
    -2.25
    -2.25
     ↵
    -2.25
    -2.25
    -2.25
    č↵
    -2.25
    <|outofrange|>
    -2.25
    <|outofrange|>
    -2.25
    <|outofrange|>
    -2.25
    POSITIVE LOGITS
     ¶
    1.31
     (,
    1.28
    ("/
    1.26
    xico
    1.25
     escal
    1.23
    ussi
    1.18
    represented
    1.15
     (;
    1.15
     denoted
    1.14
    *]{}
    1.13
    Activations Density 3.673%

    No Known Activations