© 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. 5507
    Prev
    Next
    INDEX
    Explanations

    threats or statements of intention to take legal action

    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
    150
    +0.15
    0.5%
    363
    +0.13
    0.4%
    57
    +0.12
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    150
    +0.15
    0.05
    271
    +0.13
    0.04
    730
    +0.12
    0.05
    Negative Logits
    mys
    -1.29
    gart
    -1.22
    angles
    -1.08
    mot
    -1.03
    ools
    -1.01
    area
    -1.01
    éĸ
    -1.01
    BO
    -0.98
    bel
    -0.98
    annis
    -0.96
    POSITIVE LOGITS
    lessly
    1.14
     adolesc
    1.04
    ulkan
    1.00
     threatened
    0.97
    quit
    0.95
     annihilation
    0.94
     repr
    0.94
    staking
    0.91
    ño
    0.91
     retaliation
    0.91
    Activations Density 0.822%

    No Known Activations