© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 11-RES-JB
    11. 9167
    Prev
    Next
    INDEX
    Explanations

    prepositions followed by verbs or nouns expressing control or manipulation

    oai_token-act-pair · gpt-3.5-turbo

    phrases associated with manipulative or controlling behaviors

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    DragonMagazine
    -0.74
     Registered
    -0.70
     Shot
    -0.68
    unciation
    -0.68
    âĿ
    -0.66
    rar
    -0.66
     Recorded
    -0.64
    ratulations
    -0.62
    oug
    -0.61
     Updated
    -0.61
    POSITIVE LOGITS
     achieve
    1.25
     accomplish
    1.15
     satisfy
    1.13
     appease
    1.12
     maximize
    1.12
     justify
    1.08
     perpetuate
    1.04
     compensate
    1.03
     ensure
    1.01
     survive
    1.00
    Activations Density 0.186%

    No Known Activations