© 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. 12884
    Prev
    Next
    INDEX
    Explanations

    contradictory or unexpected statements

    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
    397
    +0.25
    1.5%
    25
    +0.15
    0.9%
    2
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    130
    +0.25
    0.10
    286
    +0.15
    0.11
    134
    +0.13
    0.11
    Negative Logits
    ¿
    -4.46
    Ń
    -4.42
    ·¸
    -4.40
    ³
    -4.36
    ¼
    -4.34
    ¾
    -4.30
    ĭ
    -4.22
    Ĩ
    -4.17
    ij
    -4.16
    Ģ
    -4.15
    POSITIVE LOGITS
     thanks
    2.62
     though
    2.58
     especially
    2.50
     but
    2.43
    though
    2.36
     thank
    2.35
    although
    2.27
    but
    2.18
     however
    2.17
     anyway
    2.17
    Activations Density 1.236%

    No Known Activations