© 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. 3857
    Prev
    Next
    INDEX
    Explanations

    phrases related to making a point or clarification

    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
    239
    +0.17
    0.5%
    682
    +0.15
    0.5%
    178
    +0.14
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    671
    +0.17
    0.04
    539
    +0.15
    0.04
    507
    +0.14
    0.04
    Negative Logits
    utenberg
    -1.36
    bah
    -1.09
    icking
    -1.06
    ufact
    -1.05
    mun
    -1.04
     Flavoring
    -1.00
    oven
    -1.00
    abb
    -0.99
    odan
    -0.99
    ngth
    -0.97
    POSITIVE LOGITS
     anymore
    1.40
     anything
    1.33
     necessarily
    1.28
    erest
    1.21
     anyone
    1.10
     any
    1.08
     anybody
    1.07
     signs
    1.05
     disrespect
    1.04
     exactly
    1.03
    Activations Density 0.467%

    No Known Activations