{"event":{"id":"evt-redpajama-20230417","dedupe_key":"open_training_data_release:2023-04-17:redpajama-releases-open-llm-training-dataset","event_type":"open_training_data_release","title":"RedPajama releases open LLM training dataset","summary":"An open 1.2T-token dataset intended to reproduce an important part of the LLaMA training-data stack.","occurred_at":"2023-04-17T00:00:00.000Z","published_at":"2023-04-17T00:00:00.000Z","observed_at":"2026-10-07T12:39:09.000Z","reconstructed_at":"2026-10-07T12:55:58.000Z","ingest_type":"live_world_scan","locations":null,"status":"verified","confidence":0.98,"metadata":{"actors":["Together and collaborators"],"affected_layers":["layer-open-source","layer-models","layer-research"],"change_kind":"open_foundation_stack_maturation","what_happened_at_time":"An open 1.2T-token dataset intended to reproduce an important part of the LLaMA training-data stack.","evidence_at_time":"The contemporaneous primary source 'RedPajama: an Open Dataset for Training Large Language Models' documents this Q2 event.","q1_continuity":"Q1 is the canonical predecessor; this record stores only the supported Q2 delta.","retrospective_inference":"Interpretation is limited to evidence supportable by 2023-06-30.","uncertainty":"Open data does not make frontier training compute broadly available."},"created_at":"2026-10-07T14:04:48.591Z","updated_at":"2026-10-07T14:04:48.591Z"},"artifacts":[{"id":"art-redpajama-20230417","dedupe_key":"b1bf20e9e8077a77907f251f9f3301687741975b5d9a382aeb136aca819760d2","canonical_url":"https://www.together.ai/blog/redpajama","evidence_locator":"","artifact_type":"open_training_dataset_release","title":"RedPajama 1.2T-token dataset","summary":"An open 1.2T-token dataset intended to reproduce an important part of the LLaMA training-data stack.","creator_entities":["Together and collaborators"],"released_at":"2023-04-17T00:00:00.000Z","content_hash":null,"metadata":{"historical_scope":"2023-Q2"},"created_at":"2026-10-07T14:04:48.584Z","updated_at":"2026-10-07T16:27:14.256Z","relation":"evidenced_by"}],"signals":[{"id":"sig-q2-open-foundation-stack","statement":"Open training data, commercially usable model weights and multiple independent foundation-model families formed a more complete reproducible/open alternative model-development stack.","signal_type":"open_foundation_stack_maturation","direction":"increasing","confidence":0.97,"epistemic_status":"supported_inference","mapping_mode":"retrospective","reconstructed_at":"2026-10-07T12:55:58.000Z","created_at":"2026-10-07T14:04:48.604Z","updated_at":"2026-10-07T14:04:48.604Z"}]}