We lifted 8 functions out of this paper's own repositories and ran 4 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 |
|---|---|---|
| meowpass/followcomplexinstruction | canonical | 4 of 8 |
| Function | Status | Where it lives |
|---|---|---|
| conflict_make | Ran | meowpass/followcomplexinstruction/get_data/instructions_registry.py pointer only (licence: NONE) · get_code("d197da35ee544f75") |
| count_words | Ran | meowpass/followcomplexinstruction/get_data/instructions_util.py pointer only (licence: NONE) · get_code("cdcc85ca09b00f7e") |
| read_prompt_list | Ran | meowpass/followcomplexinstruction/get_data/check.py pointer only (licence: NONE) · get_code("6ef1aee592402546") |
| split_into_sentences | Ran | meowpass/followcomplexinstruction/get_data/instructions_util.py pointer only (licence: NONE) · get_code("34d346853108011e") |
| count_sentences | Not yet run | meowpass/followcomplexinstruction/get_data/instructions_util.py pointer only (licence: NONE) · get_code("f79646ddd0a03162") |
| format_prompt | Not yet run | meowpass/followcomplexinstruction/get_data/do_inference.py pointer only (licence: NONE) · get_code("ee0fb943a653c769") |
| generate | Not yet run | meowpass/followcomplexinstruction/get_data/do_inference.py pointer only (licence: NONE) · get_code("44bfe81032401eb3") |
| get_res | Not yet run | meowpass/followcomplexinstruction/get_data/GPT_api.py pointer only (licence: NONE) · get_code("6c1a7db24cb308eb") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
It is imperative for Large language models (LLMs) to follow instructions with elaborate requirements (i.e. Complex Instructions Following). Yet, it remains under-explored how to enhance the ability of LLMs to follow complex instructions with multiple constraints. To bridge the gap, we initially study what training data is effective in enhancing complex constraints following abilities. We found that training LLMs with instructions containing multiple constraints enhances their understanding of complex instructions, especially those with lower complexity levels. The improvement can even generalize to compositions of out-of-domain constraints. Additionally, we further propose methods addressing how to obtain and utilize the effective training data. Finally, we conduct extensive experiments to prove the effectiveness of our methods in terms of overall performance and training efficiency. We also demonstrate that our methods improve models' ability to follow instructions generally and generalize effectively across out-of-domain, in-domain, and adversarial settings, while maintaining general capabilities.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2404.15846")
get_code_for_paper("2404.15846")
have("2404.15846")
Connect an agent — have() is free.