© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    5. GPT2-Small
    7. Residual Similar CE E2E Recon
    9. 6-RES_SCEFR-AJT
    11. 13665
    Prev
    Next
    INDEX
    Explanations

    references to a specific person or proper noun

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-ajt
    Prompts (Dashboard)
    12,288 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    46,080
    Data Type
    torch.float32
    Hook Point
    blocks.6.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    apollo-research/Skylion007-openwebtext-tokenizer-gpt2
    Hook Point Layer
    6
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    146
    +0.08
    0.5%
    369
    +0.07
    0.5%
    46
    +0.07
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    154
    +0.08
    0.06
    514
    +0.07
    0.04
    23
    +0.07
    0.04
    Negative Logits
    FG
    -0.55
     disg
    -0.55
    IRD
    -0.52
    film
    -0.51
    Reviewed
    -0.49
    gged
    -0.49
    RET
    -0.47
    hold
    -0.46
    erent
    -0.45
    DIV
    -0.45
    POSITIVE LOGITS
    idate
    0.55
     Commands
    0.55
    uyomi
    0.55
    rity
    0.53
     Clause
    0.52
    alker
    0.51
    ciating
    0.51
    atism
    0.50
    ient
    0.50
    racuse
    0.49
    Activations Density 0.085%

    No Known Activations