LLM EXPLORER 59,420 MODELS INDEXED

Pythia 12B Pre V8.12.5K Steps by OpenAssistant

By OpenAssistant · 76 downloads

Pythia 12B Pre V8.12.5K Steps is an open-source language model by OpenAssistant. Features: 12b LLM, VRAM: 23.8GB, Context: 2K, License: apache-2.0, LLM Explorer Score: 0.1, Arc: 41.5, HellaSwag: 68.8, MMLU: 26.6, GSM8K: 7.7.

  Endpoints compatible   Gpt neox   Pytorch   Region:us   Sharded

Pythia 12B Pre V8.12.5K Steps Parameters and Internals

Model Type 
text generation
Additional Notes 
This is an intermediate model used as a base for further experiments, not intended for use.
Training Details 
Data Sources:
gpteacher_roleplay, red_pajama, wizardlm_70k, joke, poem_instructions, oa_stackexchange, tell_a_joke, webgpt, gpt4all, alpaca_gpt4, code_alpaca, vicuna, oig_file, minimath, humaneval_mbpp_codegen_qa, humaneval_mbpp_testgen_qa, grade_school_math_instructions, recipes, cmu_wiki_qa, oa_wiki_qa_bart_10000row, prosocial_dialogue, explain_prosocial, soda, oa_leet10k, dolly15k
Data Volume:
12.5k steps with batch-size 128
Methodology:
Instruction-tuning
Context Length:
2048
Model Architecture:
pythia-12b-pre-v8
LLM NamePythia 12B Pre V8.12.5K Steps
Repository πŸ€—https://huggingface.co/OpenAssistant/pythia-12b-pre-v8-12.5k-steps 
Model Size12b
Required VRAM23.8 GB
Updated2026-07-28
MaintainerOpenAssistant
Model Typegpt_neox
Model Files  10.0 GB: 1-of-3   9.9 GB: 2-of-3   3.9 GB: 3-of-3
Model ArchitectureGPTNeoXForCausalLM
Licenseapache-2.0
Context Length2048
Model Max Length2048
Transformers Version4.28.0.dev0
Tokenizer ClassGPTNeoXTokenizer
Vocabulary Size50288
Torch Data Typefloat16

Best Alternatives to Pythia 12B Pre V8.12.5K Steps

Best Alternatives
Context / RAM
Downloads
Likes
Dolly V2 12B2K / 23.8 GB32831955
...sst Sft 4 Pythia 12B Epoch 3.52K / 23.8 GB818368
Pythia 12B2K / 23.8 GB147000145
Oasst Rl 1 Pythia 12B2K / 23.8 GB104
...st Sft 3 Pythia 12B Epoch 2.352K / 23.8 GB100
Pythia 12B Deduped2K / 23.8 GB417253
Oasst Sft 1 Pythia 12B2K / 23.8 GB301277
Pythia 12B Sft V8 7K Steps2K / 23.8 GB110622
Oasst Pythia 12B Reference2K / 23.8 GB14420
H2ogpt Oasst1 512 12B2K / 23.9 GB24229
Note: green Score (e.g. "73.2") means that the model is better than OpenAssistant/pythia-12b-pre-v8-12.5k-steps.