1experiment_id: E1_core2splits:3- test4families:5- synthetic6- exec7lengths:8- 20009- 800010- 1600011limit_docs_per_family: 4012max_questions_per_doc: 513model_pairs:14- sender: llama3b15 receiver: llama3b16generation:17 max_tokens: 4818 per_family:19 exec: 32020seeds:21- 022conditions:23- method: A824 budget_tokens: 51225 access: blind26note: 'Retry of A8 (LLMLingua-2) for the Llama pair on synthetic+exec: the M3U96a27 environment lacked torch (implementation_error); rerun on the laptop, new attempts28 appended to E1_core'29protocol_hash: 404c52b34d964693da72ef459a036ffc42fbf5f6ba8748cc9de0c554ebebddd630