brightlabYear 12 · Recorded LLM traces + native rerun · 55 minutes
Inside a pretrained language model
Can restoring an internal state recover a language model’s answer preference?
The first load downloads Python and may take a moment. Write a prediction, then open the lab. Controls, help, teacher notes and your evidence download are inside. If the school network blocks the runtime, download the Python notebook and run it with your teacher.
Investigate one research connection
Mixing Mechanisms: How Language Models Retrieve Bound Entities In-Context
Patch clean activations into a corrupted binding prompt, then compare multiple prompts. Record the source model/task, a baseline, the changed factor, a measurement and an alternative explanation. State exactly which part your notebook investigates and which part it does not reproduce.
Our tiny binding transformer is independently trained and does not reproduce the paper's nine-model circuit findings.
Use the notebook below to collect the measurements. This writing remains in this page session until downloaded.
Teacher background and source method
Llama, Gemma and Qwen families, 2–72B, ten binding tasks
Uses ablations to separate positional, lexical and reflexive retrieval; compares causal predictions as entity lists and context length change.
Independent classroom adaptation; not a reproduction of the source model or complete method.
Read the source with a teacher ↗Opening the notebook page…
First use downloads Python and its libraries. A fresh session measured about 20–22 seconds and 14–17 MB during the initial audit; school networks vary. This hosted notebook computes on the browser CPU.
Your prediction and saved runs stay in this session. Download your evidence before leaving or refreshing. The browser explores recorded CPU measurements. Repeating the model experiment needs native Python; CPU is sufficient and an H100 is optional.
Research connections · 3 archive entries
These lessons adapt ideas and methods. The original sources state their own model, data and validation scope.
Mixing Mechanisms: How Language Models Retrieve Bound Entities In-Context ↗
Our tiny binding transformer is independently trained and does not reproduce the paper's nine-model circuit findings.
Replicating Circuit Tracing for a Simple Known Mechanism ↗
Replication found both similarities and differences. Our attention patching does not reproduce CLT circuit discovery.
The Circuits Research Landscape: Results and Perspectives ↗
The linked community guide was reviewed. Attribution graphs provide hypotheses with missing or uninterpretable components; interventions remain necessary.
Pythia model card · Raw measurements and provenance ↓ · H100 setup guide