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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar CE Local
    5. 6-RES_SCL-AJT
    6. 34168
    Prev
    Next
    INDEX
    Explanations

    phrases related to returning or being back

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-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
    447
    +0.14
    0.7%
    760
    +0.09
    0.5%
    372
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    223
    +0.14
    0.08
    197
    +0.09
    0.07
    161
    +0.08
    0.04
    Negative Logits
    uously
    -0.58
    kefeller
    -0.56
    anguard
    -0.52
    entric
    -0.52
     Hots
    -0.52
    "]=>
    -0.52
     delegation
    -0.52
     Institution
    -0.51
     JPM
    -0.51
    ision
    -0.50
    POSITIVE LOGITS
    tracking
    0.92
    dated
    0.85
    GROUND
    0.82
    fires
    0.81
    packs
    0.78
    packing
    0.75
    door
    0.75
    ported
    0.72
    draft
    0.72
    ordered
    0.71
    Activations Density 5.970%

    No Known Activations