We lifted 13 functions out of this paper's own repositories and ran 9 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| RyanLiu112/compute-optimal-tts | canonical | 9 of 13 |
| Function | Status | Where it lives |
|---|---|---|
| choice_answer_clean | Ran | RyanLiu112/compute-optimal-tts/src/envs/MATH/grader.py code served (permissive licence) · get_code("809c0178a7ae90ff") |
| count_unknown_letters_in_expr | Ran | RyanLiu112/compute-optimal-tts/src/envs/MATH/verify_utils.py code served (permissive licence) · get_code("368d3db6347269c3") |
| get_model_name | Ran | RyanLiu112/compute-optimal-tts/src/utils.py code served (permissive licence) · get_code("5e9f6ab031880280") |
| get_root | Ran | RyanLiu112/compute-optimal-tts/src/reason/guided_search/tree.py code served (permissive licence) · get_code("105b6f113614fb89") |
| is_digit | Ran | RyanLiu112/compute-optimal-tts/src/envs/MATH/grader.py code served (permissive licence) · get_code("fb37d3c84b18e368") |
| is_file_exists | Ran | RyanLiu112/compute-optimal-tts/src/utils.py code served (permissive licence) · get_code("6d13ee4c8c032e97") |
| normalize_answer | Ran | RyanLiu112/compute-optimal-tts/src/envs/MATH/verify_utils.py code served (permissive licence) · get_code("101b1e54fcdadf90") |
| parse_digits | Ran | RyanLiu112/compute-optimal-tts/src/envs/MATH/grader.py code served (permissive licence) · get_code("7d4b7f8c3875b126") |
| should_allow_eval | Ran | RyanLiu112/compute-optimal-tts/src/envs/MATH/verify_utils.py code served (permissive licence) · get_code("95f02454eddf082a") |
| extract_answer | Not yet run | RyanLiu112/compute-optimal-tts/src/envs/MATH/env.py code served (permissive licence) · get_code("042094830d54a879") |
| extract_groundtruth | Not yet run | RyanLiu112/compute-optimal-tts/src/envs/MATH/env.py code served (permissive licence) · get_code("664b4759319bf405") |
| is_file_empty | Not yet run | RyanLiu112/compute-optimal-tts/src/utils.py code served (permissive licence) · get_code("c47dcad7c32e67e8") |
| judge_correct | Not yet run | RyanLiu112/compute-optimal-tts/src/envs/MATH/env.py code served (permissive licence) · get_code("25655b21b478adb7") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Test-Time Scaling (TTS) is an important method for improving the performance of Large Language Models (LLMs) by using additional computation during the inference phase. However, current studies do not systematically analyze how policy models, Process Reward Models (PRMs), and problem difficulty influence TTS. This lack of analysis limits the understanding and practical use of TTS methods. In this paper, we focus on two core questions: (1) What is the optimal approach to scale test-time computation across different policy models, PRMs, and problem difficulty levels? (2) To what extent can extended computation improve the performance of LLMs on complex tasks, and can smaller language models outperform larger ones through this approach? Through comprehensive experiments on MATH-500 and challenging AIME24 tasks, we have the following observations: (1) The compute-optimal TTS strategy is highly dependent on the choice of policy model, PRM, and problem difficulty. (2) With our compute-optimal TTS strategy, extremely small policy models can outperform larger models. For example, a 1B LLM can exceed a 405B LLM on MATH-500. Moreover, on both MATH-500 and AIME24, a 0.5B LLM outperforms GPT-4o, a 3B LLM surpasses a 405B LLM, and a 7B LLM beats o1 and DeepSeek-R1, while with higher inference efficiency. These findings show the significance of adapting TTS strategies to the specific characteristics of each task and model and indicate that TTS is a promising approach for enhancing the reasoning abilities of LLMs.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2502.06703")
get_code_for_paper("2502.06703")
have("2502.06703")
Connect an agent — have() is free.