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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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 E2E Recon
    5. 6-RES_SCEFR-AJT
    6. 730
    Prev
    Next
    INDEX
    Explanations

    phrases related to rewards and motivations

    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
    288
    +0.19
    0.6%
    620
    +0.18
    0.6%
    35
    +0.14
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.19
    0.09
    453
    +0.18
    0.05
    424
    +0.14
    0.04
    Negative Logits
    aug
    -1.07
    enium
    -1.02
    okin
    -1.00
    gow
    -0.99
    obiles
    -0.97
    uka
    -0.96
    enic
    -0.95
    ovies
    -0.95
    soType
    -0.93
    pac
    -0.93
    POSITIVE LOGITS
     rewarded
    1.28
     reward
    1.26
     rewards
    1.14
     reinforcement
    1.05
     payoff
    1.02
     reap
    0.99
     punishment
    0.99
     sanction
    0.97
     punish
    0.96
     accordingly
    0.96
    Activations Density 1.103%

    No Known Activations