DefCPM-2B aims to be a lightweight truth-seeking sycophancy-free LLM assistant.
After experimenting with several LLMs (some of which I have shared on this account), I came to the conclusion that many Instruction-Tuned models would rather sacrifice accuracy that to dissatisfy the user (a phenomenon known as sycophancy).
Therefore, I created
this dataset
to tackle with this exact problem, to encourage LLMs to give precise answers, teaching them to directly contest false user premises.
This model was fine-tuned (3 epochs, 0,00001 learning rate) with said dataset as an application of the concept. It adopted the expected behavior surprisingly well.
I generally recommend a Q4_K_M or an IQ4_XS quantization for deployment on edge devices.
Enjoy!
DefCPM stands for 'Definitive CPM'.
Runs of aiuser3993 DefCPM-2B on huggingface.co
449
Total runs
2
24-hour runs
8
3-day runs
40
7-day runs
140
30-day runs
More Information About DefCPM-2B huggingface.co Model
DefCPM-2B huggingface.co is an AI model on huggingface.co that provides DefCPM-2B's model effect (), which can be used instantly with this aiuser3993 DefCPM-2B model. huggingface.co supports a free trial of the DefCPM-2B model, and also provides paid use of the DefCPM-2B. Support call DefCPM-2B model through api, including Node.js, Python, http.
DefCPM-2B huggingface.co is an online trial and call api platform, which integrates DefCPM-2B's modeling effects, including api services, and provides a free online trial of DefCPM-2B, you can try DefCPM-2B online for free by clicking the link below.
aiuser3993 DefCPM-2B online free url in huggingface.co:
DefCPM-2B is an open source model from GitHub that offers a free installation service, and any user can find DefCPM-2B on GitHub to install. At the same time, huggingface.co provides the effect of DefCPM-2B install, users can directly use DefCPM-2B installed effect in huggingface.co for debugging and trial. It also supports api for free installation.