Update README.md
Browse files
README.md
CHANGED
@@ -13,15 +13,14 @@ base_model:
|
|
13 |
This is a LLM fine-tuned with real-life, ideally-empathetic teacher-student conversations.
|
14 |
This model processes the recent conversation history and provides guidance on how a teacher might respond to the student's utterance.
|
15 |
|
16 |
-
To fine-tune an open-weighted LLM to act as this generic teacher, we
|
17 |
-
the Teacher-Student Chatroom Corpus,
|
18 |
-
TSCCv2 [Caines et al., 2022](https://aclanthology.org/2022.nlp4call-1.3),
|
19 |
CIMA [Stasaski et al., 2020](https://aclanthology.org/2020.bea-1.5),
|
20 |
the Multicultural Classroom Discourse Dataset [Rapanta et al., 2021](https://www.sciencedirect.com/science/article/pii/S2352340921007940),
|
21 |
MathDial [Macina et al., 2023](https://aclanthology.org/2023.findings-emnlp.372), and
|
22 |
Conversational Uptake [Demszky et al., 2021].
|
23 |
|
24 |
-
We are evaluating LLaMa-3
|
25 |
Instead of using programmable fine-tuning libraries such as Axolotl ([link](https://github.com/OpenAccess-AI-Collective/axolotl))
|
26 |
or Huggingface TRL ([link](https://github.com/huggingface/trl)),
|
27 |
we are employing the more general command-line LLaMA-Factory ([link](https://github.com/hiyouga/LLaMA-Factory)) toolkit
|
|
|
13 |
This is a LLM fine-tuned with real-life, ideally-empathetic teacher-student conversations.
|
14 |
This model processes the recent conversation history and provides guidance on how a teacher might respond to the student's utterance.
|
15 |
|
16 |
+
To fine-tune an open-weighted LLM to act as this generic teacher, we have used the following datasets:
|
17 |
+
the Teacher-Student Chatroom Corpus, TSCCv2 [Caines et al., 2022](https://aclanthology.org/2022.nlp4call-1.3),
|
|
|
18 |
CIMA [Stasaski et al., 2020](https://aclanthology.org/2020.bea-1.5),
|
19 |
the Multicultural Classroom Discourse Dataset [Rapanta et al., 2021](https://www.sciencedirect.com/science/article/pii/S2352340921007940),
|
20 |
MathDial [Macina et al., 2023](https://aclanthology.org/2023.findings-emnlp.372), and
|
21 |
Conversational Uptake [Demszky et al., 2021].
|
22 |
|
23 |
+
We are evaluating LLaMa-3 for this task.
|
24 |
Instead of using programmable fine-tuning libraries such as Axolotl ([link](https://github.com/OpenAccess-AI-Collective/axolotl))
|
25 |
or Huggingface TRL ([link](https://github.com/huggingface/trl)),
|
26 |
we are employing the more general command-line LLaMA-Factory ([link](https://github.com/hiyouga/LLaMA-Factory)) toolkit
|