© 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. 11050
    Prev
    Next
    INDEX
    Explanations

    phrases and terms associated with legal contexts and compliance issues

    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
    21
    +0.24
    1.6%
    156
    +0.21
    1.4%
    78
    +0.12
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    21
    +0.24
    0.06
    261
    +0.21
    0.04
    37
    +0.12
    0.00
    Negative Logits
    ger
    -1.41
    idation
    -1.39
    ĥ½
    -1.39
    ĨĴ
    -1.34
    irc
    -1.29
      ↵
    -1.27
    ↵
    -1.27
    <|outofrange|>
    -1.27
    ↵
    -1.27
    ↵↵
    -1.27
    POSITIVE LOGITS
     Himself
    1.57
     contradiction
    1.47
    himself
    1.40
    circum
    1.38
     His
    1.25
     bishop
    1.25
     encephal
    1.24
     assumption
    1.21
    Appeal
    1.21
    unto
    1.20
    Activations Density 0.777%

    No Known Activations