diff --git a/configs/llm-360-amber1.yaml b/configs/llm-360-amber1.yaml
new file mode 100644
index 000000000..17951ce85
--- /dev/null
+++ b/configs/llm-360-amber1.yaml
@@ -0,0 +1,598 @@
+run_name: amberish1-base
+seed: 6198
+dry_run: false
+
+wandb:
+  name: ${run_name}
+  project: olmo-small
+  group: amberish1-base
+
+model:
+  d_model: 2048
+  n_heads: 16
+  n_layers: 16
+  mlp_ratio: 8
+  # mlp_hidden_size: 22016
+  weight_tying: false
+  alibi: false
+  rope: true
+  flash_attention: true
+  attention_dropout: 0.0
+  attention_layer_norm: false
+  clip_qkv: null
+  include_bias: false
+  block_type: sequential
+  layer_norm_type: rms
+  layer_norm_with_affine: false
+  layer_norm_eps: 1e-6
+  bias_for_layer_norm: false
+  attention_layer_norm_with_affine: false
+  activation_type: swiglu
+  residual_dropout: 0.0
+  embedding_dropout: 0.0
+  max_sequence_length: 2048
+  vocab_size: 32000
+  embedding_size: 32000
+  eos_token_id: 2
+  pad_token_id: 2
+  init_device: meta
+  init_fn: normal
+  init_std: 0.02
+  init_cutoff_factor: 3
+
+compile: null
+
+optimizer:
+  name: adamw
+  learning_rate: 4.0e-4
+  weight_decay: 0.1
+  decay_norm_and_bias: true
+  decay_embeddings: false
+  eps: 1.0E-08
+  betas:
+  - 0.9
+  - 0.95
+  metrics_log_interval: 10
+
+scheduler:
+  name: cosine_with_warmup
+  units: tokens
+  t_warmup: 4587520000
+  t_max: 1.25e12
+  alpha_f: 0.1
+  warmup_min_lr: 0
+
+tokenizer:
+  identifier: huggyllama/llama-7b
+  truncate_direction: right
+
+save_folder:  /weka/oe-training-default/oe-training-default/ai2-llm/checkpoints/OLMo-medium/dustins-stability/${run_name}
+# remote_save_folder: /weka/oe-training-default/ai2-llm/checkpoints/OLMo-small/${run_name}
+save_overwrite: true
+
+save_interval: 1000
+save_interval_ephemeral: null
+save_num_checkpoints_to_keep: -1
+sharded_checkpointer: olmo_core
+
+save_interval_unsharded: null
+save_num_unsharded_checkpoints_to_keep: -1
+
+load_path: null
+
+max_duration: 2ep
+global_train_batch_size: 1120
+device_train_microbatch_size: 4
+
+precision: amp_bf16
+
+fsdp:
+  wrapping_strategy: by_block_and_size
+  sharding_strategy: SHARD_GRAD_OP
+  precision: mixed
+
+max_grad_norm: 1.0
+max_grad_norm_ratio: null
+
+speed_monitor:
+  window_size: 1
+
+eval_interval: 1000
+eval_subset_num_batches: -1
+device_eval_batch_size: ${device_train_microbatch_size}
+evaluators:
+  # - label: all-small-ppl-validation
+  #   data:
+  #     num_workers: 0
+  #     drop_last: true
+  #     datasets:
+  #       c4_en-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/c4_en/val/part-0-00000.npy
+  #       dolma_books-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/dolma_books/val/part-0-00000.npy
+  #       dolma_common-crawl-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/dolma_common-crawl/val/part-0-00000.npy
+  #       dolma_pes2o-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/dolma_pes2o/val/part-0-00000.npy
+  #       dolma_reddit-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/dolma_reddit/val/part-0-00000.npy
+  #       dolma_stack-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/dolma_stack/val/part-0-00000.npy
+  #       dolma_wiki-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/dolma_wiki/val/part-0-00000.npy
+  #       ice-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/ice/val/part-0-00000.npy
+  #       m2d2_s2orc-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/m2d2_s2orc/val/part-0-00000.npy
+  #       pile-validation:
+  #       - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/pile/val/part-0-00000.npy
+  #       wikitext_103-validation:
+  #         - /weka/oe-training-default/ai2-llm/eval-data/perplexity/v3_small_gptneox20b/wikitext_103/val/part-0-00000.npy
+
+  ##########################
+  # Downstream evaluations #
+  ##########################
+  - label: piqa
+    type: downstream
+
+  - label: hellaswag
+    type: downstream
+
+  - label: winogrande
+    type: downstream
+
+  - label: openbook_qa
+    type: downstream
+
+  - label: boolq
+    type: downstream
+ 
+  - label: sciq
+    type: downstream
+
+  - label: arc_easy
+    type: downstream
+
+  - label: arc_challenge
+    type: downstream
+
+  - label: copa
+    type: downstream
+
+  #- label: rte
+  #  type: downstream
+
+  #- label: commitment_bank
+  #  type: downstream
+
+  #- label: sst2
+  #  type: downstream
+
+  - label: commonsense_qa
+    type: downstream
+
+  - label: social_iqa
+    type: downstream
+
+  # Doesn't work from cache.
+  # - label: basic_arithmetic
+  #   type: downstream
+
+  - label: mmlu_stem_var
+    type: downstream
+
+  - label: mmlu_humanities_var
+    type: downstream
+
+  - label: mmlu_social_sciences_var
+    type: downstream
+
+  - label: mmlu_other_var
+    type: downstream
+
+  - label: mmlu_stem_mc_5shot
+    type: downstream
+
+  - label: mmlu_humanities_mc_5shot
+    type: downstream
+
+  - label: mmlu_social_sciences_mc_5shot
+    type: downstream
+
+  - label: mmlu_other_mc_5shot
+    type: downstream
+
+  - label: mmlu_stem_mc_5shot_test
+    type: downstream
+
+  - label: mmlu_humanities_mc_5shot_test
+    type: downstream
+
+  - label: mmlu_social_sciences_mc_5shot_test
+    type: downstream
+
+  - label: mmlu_other_mc_5shot_test
+    type: downstream
+
+  - label: basic_arithmetic
+    type: downstream
+
+  - label: trivia_qa_wiki_ppl
+    type: downstream
+
+  - label: natural_qs_open_ppl
+    type: downstream
+
+  - label: arc_easy_ppl
+    type: downstream
+
+data:
+  pad_direction: right
+  num_workers: 32
+  drop_last: true
+  pin_memory: true
+  prefetch_factor: 8
+  persistent_workers: true
+  timeout: 0
+  paths:
+    ######### Amber #########
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_000.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_001.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_002.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_003.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_004.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_005.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_006.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_007.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_008.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_009.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_010.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_011.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_012.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_013.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_014.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_015.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_016.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_017.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_018.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_019.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_020.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_021.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_022.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_023.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_024.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_025.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_026.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_027.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_028.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_029.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_030.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_031.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_032.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_033.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_034.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_035.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_036.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_037.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_038.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_039.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_040.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_041.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_042.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_043.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_044.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_045.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_046.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_047.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_048.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_049.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_050.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_051.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_052.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_053.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_054.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_055.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_056.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_057.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_058.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_059.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_060.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_061.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_062.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_063.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_064.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_065.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_066.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_067.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_068.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_069.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_070.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_071.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_072.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_073.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_074.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_075.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_076.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_077.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_078.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_079.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_080.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_081.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_082.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_083.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_084.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_085.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_086.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_087.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_088.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_089.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_090.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_091.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_092.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_093.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_094.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_095.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_096.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_097.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_098.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_099.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_100.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_101.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_102.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_103.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_104.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_105.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_106.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_107.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_108.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_109.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_110.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_111.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_112.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_113.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_114.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_115.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_116.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_117.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_118.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_119.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_120.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_121.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_122.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_123.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_124.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_125.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_126.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_127.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_128.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_129.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_130.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_131.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_132.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_133.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_134.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_135.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_136.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_137.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_138.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_139.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_140.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_141.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_142.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_143.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_144.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_145.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_146.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_147.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_148.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_149.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_150.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_151.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_152.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_153.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_154.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_155.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_156.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_157.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_158.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_159.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_160.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_161.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_162.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_163.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_164.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_165.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_166.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_167.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_168.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_169.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_170.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_171.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_172.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_173.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_174.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_175.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_176.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_177.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_178.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_179.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_180.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_181.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_182.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_183.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_184.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_185.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_186.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_187.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_188.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_189.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_190.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_191.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_192.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_193.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_194.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_195.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_196.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_197.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_198.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_199.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_200.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_201.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_202.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_203.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_204.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_205.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_206.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_207.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_208.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_209.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_210.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_211.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_212.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_213.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_214.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_215.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_216.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_217.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_218.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_219.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_220.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_221.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_222.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_223.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_224.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_225.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_226.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_227.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_228.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_229.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_230.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_231.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_232.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_233.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_234.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_235.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_236.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_237.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_238.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_239.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_240.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_241.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_242.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_243.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_244.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_245.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_246.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_247.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_248.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_249.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_250.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_251.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_252.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_253.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_254.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_255.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_256.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_257.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_258.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_259.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_260.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_261.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_262.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_263.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_264.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_265.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_266.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_267.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_268.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_269.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_270.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_271.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_272.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_273.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_274.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_275.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_276.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_277.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_278.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_279.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_280.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_281.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_282.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_283.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_284.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_285.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_286.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_287.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_288.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_289.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_290.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_291.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_292.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_293.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_294.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_295.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_296.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_297.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_298.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_299.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_300.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_301.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_302.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_303.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_304.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_305.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_306.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_307.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_308.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_309.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_310.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_311.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_312.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_313.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_314.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_315.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_316.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_317.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_318.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_319.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_320.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_321.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_322.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_323.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_324.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_325.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_326.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_327.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_328.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_329.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_330.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_331.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_332.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_333.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_334.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_335.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_336.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_337.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_338.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_339.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_340.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_341.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_342.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_343.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_344.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_345.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_346.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_347.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_348.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_349.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_350.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_351.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_352.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_353.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_354.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_355.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_356.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_357.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_358.npy
+    - /weka/oe-training-default/ai2-llm/preprocessed/llm-360-amber-datasets/olmo_npy/train_359.npy
\ No newline at end of file
diff --git a/olmo/config.py b/olmo/config.py
index 7c294b2db..96c761899 100644
--- a/olmo/config.py
+++ b/olmo/config.py
@@ -349,6 +349,8 @@ class ModelConfig(BaseConfig):
     to ``False``.
     """
 
+    layer_norm_eps: float = 1e-05
+
     attention_layer_norm_with_affine: bool = True
     """
     Toggle affine transform for the QK norms.
diff --git a/olmo/model.py b/olmo/model.py
index 65c430e78..f902c2463 100644
--- a/olmo/model.py
+++ b/olmo/model.py
@@ -136,11 +136,10 @@ def __init__(
         *,
         size: Optional[int] = None,
         elementwise_affine: Optional[bool] = True,
-        eps: float = 1e-05,
     ):
         super().__init__()
         self.config = config
-        self.eps = eps
+        self.eps = config.layer_norm_eps
         self.normalized_shape = (size or config.d_model,)
         if elementwise_affine or (elementwise_affine is None and self.config.layer_norm_with_affine):
             self.weight = nn.Parameter(torch.ones(self.normalized_shape, device=config.init_device))
@@ -199,9 +198,8 @@ def __init__(
         size: Optional[int] = None,
         low_precision: bool = False,
         elementwise_affine: Optional[bool] = None,
-        eps: float = 1e-05,
     ):
-        super().__init__(config, size=size, elementwise_affine=elementwise_affine, eps=eps)
+        super().__init__(config, size=size, elementwise_affine=elementwise_affine)
         self.low_precision = low_precision
 
     def forward(self, x: torch.Tensor) -> torch.Tensor:
@@ -230,9 +228,8 @@ def __init__(
         config: ModelConfig,
         size: Optional[int] = None,
         elementwise_affine: Optional[bool] = None,
-        eps: float = 1e-5,
     ):
-        super().__init__(config, size=size, elementwise_affine=elementwise_affine, eps=eps)
+        super().__init__(config, size=size, elementwise_affine=elementwise_affine)
 
     def forward(self, x: torch.Tensor) -> torch.Tensor:
         with torch.autocast(enabled=False, device_type=x.device.type):
diff --git a/pyproject.toml b/pyproject.toml
index 18314fc64..2d8927fd1 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -14,7 +14,7 @@ requires-python = ">=3.8"
 license = { file = "LICENSE" }
 dependencies = [
     "numpy",
-    "torch>=2.1,<2.3",
+    "torch>=2.1,<=2.3",
     "ai2-olmo-core==0.1.0",
     "omegaconf",
     "rich",
diff --git a/scripts/beaker/llamaish7-normal-launch.sh b/scripts/beaker/llamaish7-normal-launch.sh
new file mode 100755
index 000000000..4938615ac
--- /dev/null
+++ b/scripts/beaker/llamaish7-normal-launch.sh
@@ -0,0 +1,33 @@
+#!/usr/bin/env bash
+
+set -ex
+
+NUM_NODES=64
+
+gantry run \
+  --workspace ai2/OLMo-training \
+  --task-name llamaish7-normal \
+  --description "OLMo medium - 7B - Llamaish Normal" \
+  --priority urgent \
+  --preemptible \
+  --beaker-image petew/olmo-torch23-gantry \
+  --cluster ai2/jupiter-cirrascale-2 \
+  --gpus 8 \
+  --replicas "${NUM_NODES}" \
+  --leader-selection \
+  --host-networking \
+  --budget ai2/oe-training \
+  --no-nfs \
+  --propagate-failure \
+  --synchronized-start-timeout 15m \
+  --env LOG_FILTER_TYPE=local_rank0_only \
+  --env OMP_NUM_THREADS=8 \
+  --env OLMO_TASK=model \
+  --env-secret WANDB_API_KEY=AKSHITAB_WANDB_API_KEY \
+  --env-secret AWS_ACCESS_KEY_ID=AKSHITAB_AWS_ACCESS_KEY_ID \
+  --env-secret AWS_SECRET_ACCESS_KEY=AKSHITAB_AWS_SECRET_ACCESS_KEY \
+  --shared-memory 10GiB \
+  --venv base \
+  --yes \
+  --timeout=-1 \
+  -- /bin/bash -c "scripts/beaker/llamaish7-normal.sh \$BEAKER_LEADER_REPLICA_HOSTNAME ${NUM_NODES} \$BEAKER_REPLICA_RANK"
diff --git a/scripts/beaker/llm-360-amber1-launch.sh b/scripts/beaker/llm-360-amber1-launch.sh
new file mode 100755
index 000000000..3ac881088
--- /dev/null
+++ b/scripts/beaker/llm-360-amber1-launch.sh
@@ -0,0 +1,34 @@
+#!/usr/bin/env bash
+
+set -ex
+
+NUM_NODES=4
+
+gantry run \
+  --workspace ai2/OLMo-training \
+  --task-name amberish1-base \
+  --description "OLMo small - 1B - Amberish with Amber data" \
+  --priority urgent \
+  --preemptible \
+  --beaker-image petew/olmo-torch23-gantry \
+  --cluster ai2/jupiter-cirrascale-2 \
+  --gpus 8 \
+  --replicas "${NUM_NODES}" \
+  --leader-selection \
+  --host-networking \
+  --budget ai2/oe-training \
+  --no-nfs \
+  --weka oe-training-default:/weka/oe-training-default \
+  --propagate-failure \
+  --synchronized-start-timeout 20m \
+  --env LOG_FILTER_TYPE=local_rank0_only \
+  --env OMP_NUM_THREADS=8 \
+  --env OLMO_TASK=model \
+  --env-secret WANDB_API_KEY=DUSTINS_WANDB_API_KEY \
+  --env-secret AWS_ACCESS_KEY_ID=DUSTINS_AWS_ACCESS_KEY_ID \
+  --env-secret AWS_SECRET_ACCESS_KEY=DUSTINS_AWS_SECRET_ACCESS_KEY \
+  --shared-memory 10GiB \
+  --venv base \
+  --yes \
+  --timeout=-1 \
+  -- /bin/bash -c "scripts/beaker/llm-360-amber1.sh \$BEAKER_LEADER_REPLICA_HOSTNAME ${NUM_NODES} \$BEAKER_REPLICA_RANK"
diff --git a/scripts/beaker/llm-360-amber1.sh b/scripts/beaker/llm-360-amber1.sh
new file mode 100755
index 000000000..1cd2cf315
--- /dev/null
+++ b/scripts/beaker/llm-360-amber1.sh
@@ -0,0 +1,39 @@
+#!/usr/bin/env bash
+set -exuo pipefail
+IFS=$'\n\t'
+
+BEAKER_LEADER_REPLICA_HOSTNAME=$1
+shift
+
+NUM_NODES=$1
+shift
+
+BEAKER_REPLICA_RANK=$1
+shift
+
+# Warm HF cache
+mkdir -p /root/.cache
+pushd /root/.cache
+curl "https://storage.googleapis.com/hf-cache/huggingface_cache_v4.tar.gz" | tar --keep-newer-files -xzf -
+popd
+export HF_DATASETS_OFFLINE=1
+
+
+torchrun \
+  --nnodes ${NUM_NODES}:${NUM_NODES} \
+  --nproc-per-node 8 \
+  --rdzv_id=12347 \
+  --rdzv_backend=static \
+  --rdzv_endpoint=$BEAKER_LEADER_REPLICA_HOSTNAME:29400 \
+  --node_rank=$BEAKER_REPLICA_RANK \
+  --rdzv_conf="read_timeout=420" \
+  scripts/train.py \
+  configs/llm-360-amber1.yaml \
+    --gen1_gc_interval=null \
+    --save_folder=runs/ \
+    --save_interval=1000 \
+    --eval_interval=1000 \
+    --optimizer.metrics_log_interval=1 \
+    --save_overwrite \
+    --save_num_checkpoints_to_keep=3 \
+    '--load_path=s3://ai2-llm/checkpoints/OLMo-small/${run_name}/step69750/'