mhla/gpt1900-d34-22btok
Updated
GPT-1900
Pre-1900 LLMs for physics reasoning. RL models are physics-only; use the SFT model for general chat. Tune temperature (0.6-0.7).
Note Base model — pre-1900 text, 22B tokens
mhla/gpt1905-d34
Updated • 1Note Base model — pre-1905 text, 40B tokens
mhla/gpt1964-d34
UpdatedNote Base model — 1900-1964 text
mhla/gpt1900-instruct-v3-sft
Updated • 2Note Instruct v3 (full) — default chat model
mhla/gpt1900-d34-v3-sft-physics
UpdatedNote Instruct v3 (safe) — no opinions, physics focus
mhla/gpt1900-d34-contradiction-rl-v6
Reinforcement Learning • UpdatedNote Contradiction RL v6 — physics eval 0.58
mhla/gpt1900-d34-contradiction-rl-v11
Reinforcement Learning • UpdatedNote Contradiction RL v11 — BEST, physics eval 1.25
mhla/pre1900-corpus
Viewer • Updated • 2.07M • 85Note Pre-1900 English text corpus with metadata
mhla/gpt1900-physics-clm
Viewer • Updated • 336k • 22Note Physics texts for continued pretraining
mhla/gpt1900-instruct-v3-data
Updated • 24Note Instruction-tuning conversation pairs
mhla/gpt1900-contradiction-eval
Viewer • Updated • 298 • 19Note Physics contradiction evaluation problems