MNLP_M3_mcqa_model - (Qwen3-0.6B-mcqa_model_2)

This model is a fine-tuned version of unsloth/Qwen3-0.6B-Base on an unknown dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training and evaluation data

Training was done on the training splits of

• MEDMCQA (34,000 random examples with seed 42)
• MMLU_stem
• MMLU_stem_10_choices agumented
• Sciq
• Ai2 Arc
• Math_qa
• ScienceQa
• Openbookqa

Training procedure

The procedure for training was done with example of any amount of choices, for each batch size we padd the options to the biggest amount of options and example has in that batch, and from there we do the training by only grabbing the last logit form doing a feedforward on the whole prompt (question with choices) and we do cross entropy loss on this last logit with the options to choose from (so we don't do cross entyropy on the whole vocabulary we only do it on the tokens of the letters of the options (e.g. A, B, C and D)).

We also template all the training examples with 7 random templates as to make the model robuts to different types of ways one could ask an MCQA question, using different prompts can make the results vary a lot

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.04
• num_epochs: 2

Evaluation Results

The model was evaluated on a suite of Multiple Choice Question Answering (MCQA) benchmarks (on its validation and test sets repsectively for each one), and NLP4education is only the approximated 1000 question and answers given to use.

Important Note on MCQA Evals Benchmark:

The performance on these benchmarks is as follows:

First evaluation: The tests where done with this prompt (type 5):

This question assesses challenging STEM problems as found on graduate standardized tests. Carefully evaluate the options and select the correct answer.

---
[Insert Question Here]
---
[Insert Choices Here, e.g.:
A. Option 1
B. Option 2
C. Option 3
D. Option 4]
---

Your response should include the letter and the exact text of the correct choice.
Example: B. Entropy increases.
Answer:

And the teseting was done on [Letter]. [Text answer]

Benchmark	Accuracy (Acc)	Normalized Accuracy (Acc Norm)
ARC Challenge	63.90%	62.41%
ARC Easy	81.64%	77.87%
GPQA	31.92%	30.58%
Math QA	31.84%	31.11%
MCQA Evals	42.60%	38.44%
MMLU	50.94%	50.94%
MMLU Pro	15.19%	13.79%
MuSR	53.04%	51.19%
NLP4Education	44.49%	41.71%
Overall	46.17%	44.23%

Second evaluation: (type 0)

The following are multiple choice questions (with answers) about knowledge and skills in advanced master-level STEM courses.

---
*[Insert Question Here]*
---
*[Insert Choices Here, e.g.:*
*A. Option 1*
*B. Option 2*
*C. Option 3*
*D. Option 4]*
---
Answer:

And the teseting was done on [Letter]. [Text answer]

Benchmark	Accuracy (Acc)	Normalized Accuracy (Acc Norm)
ARC Challenge	67.17%	64.51%
ARC Easy	83.71%	79.57%
GPQA	28.35%	28.79%
Math QA	36.38%	34.66%
MCQA Evals	45.06%	38.31%
MMLU	50.68%	50.68%
MMLU Pro	16.22%	14.31%
MuSR	53.04%	51.19%
NLP4Education	48.71%	44.18%
Overall	47.70%	45.13%

Third evaluation: (type 2)

This is part of an assessment on graduate-level science, technology, engineering, and mathematics (STEM) concepts. Each question is multiple-choice and requires a single correct answer.

---
*[Insert Question Here]*
---
*[Insert Choices Here, e.g.:*
*A. Option 1*
*B. Option 2*
*C. Option 3*
*D. Option 4]*
---
For grading purposes, respond with: [LETTER]. [VERBATIM TEXT]
Example: D. Planck constant
Your Response:

And the teseting was done on [Letter]. [Text answer]

Benchmark	Accuracy (Acc)	Normalized Accuracy (Acc Norm)
ARC Challenge	49.97%	46.02%
ARC Easy	63.34%	55.84%
GPQA	17.41%	20.09%
Math QA	29.90%	29.50%
MCQA Evals	33.64%	32.47%
MMLU	50.94%	50.94%
MMLU Pro	14.09%	11.21%
MuSR	53.04%	51.19%
NLP4Education	38.47%	37.06%
Overall	38.98%	37.15%

First evaluation: (type 0)

The following are multiple choice questions (with answers) about knowledge and skills in advanced master-level STEM courses.

---
*[Insert Question Here]*
---
*[Insert Choices Here, e.g.:*
*A. Option 1*
*B. Option 2*
*C. Option 3*
*D. Option 4]*
---
Answer:

And the teseting was done on [Letter]

Benchmark	Accuracy (Acc)	Normalized Accuracy (Acc Norm)
ARC Challenge	68.46%	68.46%
ARC Easy	84.11%	84.11%
GPQA	37.95%	37.95%
Math QA	39.31%	39.31%
MCQA Evals	45.06%	45.06%
MMLU	50.75%	50.75%
MMLU Pro	19.25%	19.25%
MuSR	51.72%	51.72%
NLP4Education	49.80%	49.80%
Overall	49.60%	49.60%

Framework versions

• Transformers 4.52.4
• Pytorch 2.7.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.0