© 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. 1-RES-SM
    11. 27832
    Prev
    Next
    INDEX
    Explanations

    verbs related to actions and their effects

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    237
    +0.14
    0.8%
    156
    +0.14
    0.8%
    264
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    264
    +0.14
    0.08
    237
    +0.14
    0.10
    15
    +0.13
    0.09
    Negative Logits
    <|outofrange|>
    -4.16
    ↵↵
    -4.16
    ↵
    -4.16
    č↵
    -4.16
    -4.16
    ↵
    -4.16
    ↵            ↵
    -4.16
    -4.16
    <|outofrange|>
    -4.16
    ↵
    -4.16
    POSITIVE LOGITS
     besides
    1.60
     slack
    1.58
    endl
    1.56
     remotely
    1.53
     induct
    1.44
     if
    1.43
     redundant
    1.40
     ?
    1.37
     better
    1.36
     stric
    1.36
    Activations Density 0.887%

    No Known Activations