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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Initial Gemma-2B-IT SAEs
    3. Gemma-2B-IT
    4. Residual Stream
    5. 12-RES-JB
    6. 13573
    Prev
    Next
    INDEX
    Explanations

    descriptions related to the sport of hockey

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    the beginnings of sentences after periods where authors are introducing new information or statements.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @eekay

    Conjunctions, transitions, and punctuation that connect clauses in explanatory text, particularly when describing relationships between ideas, making comparisons, or providing additional context about a subject.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @eekay
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Hook Layer
    12
    Architecture
    standard
    Context Size
    1,024
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    1577
    +0.27
    1.0%
    184
    +0.15
    0.6%
    609
    +0.14
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    184
    +0.27
    0.06
    1577
    +0.15
    0.10
    163
    +0.14
    0.01
    Negative Logits
     mef
    -1.06
     wien
    -1.05
     dispen
    -1.04
     „,
    -1.03
     overla
    -1.03
     dises
    -1.03
     igno
    -1.02
     deleter
    -1.02
     abnorm
    -1.00
     effe
    -1.00
    POSITIVE LOGITS
    AndEndTag
    0.68
    minecraftforge
    0.46
     really
    0.43
    Leider
    0.43
     propOrder
    0.42
     […]
    0.41
    jsoup
    0.41
    Показать
    0.41
    Significado
    0.40
    Gibt
    0.39
    Activations Density 1.927%

    No Known Activations