© 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 Local
    5. 6-RES_SCL-AJT
    6. 39916
    Prev
    Next
    INDEX
    Explanations

    phrase fragments or identifiers related to news articles

    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
    •Layer 6 UMAP region: Local cluster above bottom line touching e2e-recon cluster - local

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    447
    +0.27
    2.0%
    679
    +0.14
    1.0%
    640
    +0.10
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.27
    0.52
    679
    +0.14
    0.40
    6
    +0.10
    0.32
    Negative Logits
     silent
    -0.46
     pup
    -0.45
     healed
    -0.42
     elbow
    -0.42
     mustache
    -0.42
     retract
    -0.41
     punched
    -0.41
     blank
    -0.41
     serpent
    -0.41
     widow
    -0.41
    POSITIVE LOGITS
     Specifically
    0.81
     Especially
    0.71
     Whereas
    0.70
     Particularly
    0.68
     Whilst
    0.68
     Previously
    0.67
     Currently
    0.66
     Firstly
    0.66
     However
    0.66
     Unfortunately
    0.65
    Activations Density 24.012%

    No Known Activations