Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 29183
    Prev
    Next
    INDEX
    Explanations

    elements related to coding structures and syntax

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    17
    +0.16
    0.9%
    156
    +0.15
    0.8%
    478
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.16
    0.02
    497
    +0.15
    0.04
    129
    +0.13
    -0.01
    Negative Logits
    hei
    -1.64
    ozo
    -1.60
    overs
    -1.59
    flash
    -1.51
    hes
    -1.50
    igu
    -1.48
    coat
    -1.46
    ól
    -1.43
    æľŁ
    -1.43
    hea
    -1.40
    POSITIVE LOGITS
     nil
    1.97
     minors
    1.84
    ®
    1.74
    ŀ
    1.74
     Bool
    1.68
     Parenthood
    1.65
    ↵       
    1.65
    ↵↵                                
    1.65
    č↵č↵   
    1.65
                                         
    1.65
    Activations Density 1.403%

    No Known Activations