© 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
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar CE E2E Recon
    5. 6-RES_SCEFR-AJT
    6. 14108
    Prev
    Next
    INDEX
    Explanations

    phrases related to high-ranking individuals, particularly in the context of official positions such as government or intelligence agencies

    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
    167
    +0.22
    0.8%
    288
    +0.19
    0.7%
    21
    +0.15
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    555
    +0.22
    0.05
    711
    +0.19
    0.04
    288
    +0.15
    0.06
    Negative Logits
    igor
    -0.98
    apesh
    -0.94
    AUT
    -0.92
     tits
    -0.88
    REG
    -0.88
    Ĥİ
    -0.87
    AST
    -0.85
    roma
    -0.85
    iky
    -0.84
     thirsty
    -0.83
    POSITIVE LOGITS
    doms
    1.21
    iate
    1.17
    ity
    1.15
     citiz
    1.11
    itiz
    1.10
    eton
    1.07
    iating
    1.05
    most
    1.03
    iates
    1.02
    itor
    0.95
    Activations Density 0.973%

    No Known Activations