This is the collection for the paper "Memorization vs. Reasoning: Updating LLMs with New Knowledge". We provide the dataset used in this paper
Aochong Oliver Li
aochongoliverli
b082f689-ebc3-413d-b1f2-27bd1e86dca7's profile picture
ganglovegang's profile picture
Ernigma's profile picture
·
AI & ML interests
Large Language Models, Natural Language Processing, Machine Learning
Recent Activity
updated
a dataset
about 2 months ago
aochongoliverli/Qwen2.5-1.5B-math8k-AM-5epochs-5e-5lr-step400-dapo-5epochs-8rollouts-16384max-len-rollouts
updated
a model
about 2 months ago
aochongoliverli/Qwen2.5-3B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step500
published
a model
about 2 months ago
aochongoliverli/Qwen2.5-3B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step500
Organizations
None yet
models
88
aochongoliverli/Qwen2.5-3B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step500
Text Generation
•
3B
•
Updated
•
1
aochongoliverli/Qwen2.5-3B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step400
Text Generation
•
3B
•
Updated
•
4
aochongoliverli/Qwen2.5-1.5B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step600
Text Generation
•
2B
•
Updated
•
2
aochongoliverli/Qwen2.5-0.5B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-5e-5lr-step500
Text Generation
•
0.5B
•
Updated
•
3
aochongoliverli/Qwen2.5-0.5B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-5e-5lr-step400
Text Generation
•
0.5B
•
Updated
•
3
aochongoliverli/Qwen2.5-7B-math8k-distill-AM-Distill-Qwen-32B-16k-10epochs-5e-5lr-step100
Updated
aochongoliverli/Qwen2.5-1.5B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step500
Text Generation
•
2B
•
Updated
•
2
aochongoliverli/Qwen2.5-1.5B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step400
Text Generation
•
2B
•
Updated
•
2
aochongoliverli/Qwen2.5-1.5B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step300
Text Generation
•
2B
•
Updated
•
2
aochongoliverli/Qwen2.5-1.5B-math8k-distill-AM-Distill-Qwen-32B-16k-5epochs-2e-5lr-step200
Text Generation
•
2B
•
Updated
•
2
datasets
67
aochongoliverli/Qwen2.5-1.5B-math8k-AM-5epochs-5e-5lr-step400-dapo-5epochs-8rollouts-16384max-len-rollouts
Viewer
•
Updated
•
7.59k
•
14
aochongoliverli/Qwen2.5-1.5B-math8k-AM-10epochs-2e-5lr-step400-dapo-5epochs-8rollouts-16384max-len-rollouts
Viewer
•
Updated
•
1.28k
•
3
aochongoliverli/Qwen2.5-0.5B-math8k-AM-400steps-dapo-5epochs-8rollouts-16384max-len-rollouts
Viewer
•
Updated
•
7.59k
•
32
aochongoliverli/Qwen2.5-1.5B-math8k-AM-400steps-dapo-5epochs-8rollouts-16384max-len-rollouts
Viewer
•
Updated
•
7.59k
•
3
aochongoliverli/Qwen4B-MegaMath-pro-max-4096-len-sft-no-external-knowledge
Viewer
•
Updated
•
2.87k
•
2
aochongoliverli/deepscaler_gptoss_high_filtered
Viewer
•
Updated
•
15k
•
24
aochongoliverli/math8k-sft-QwQ-32B-16k-reasoning-traces-limo1000-len-4096
Viewer
•
Updated
•
1.08k
•
5
aochongoliverli/deepscaler_gptoss_high
Viewer
•
Updated
•
35.5k
•
4
aochongoliverli/Qwen2.5-3B-math8k-AM-400steps-dapo-5epochs-8rollouts-16384max-len-rollouts
Viewer
•
Updated
•
7.59k
•
24
aochongoliverli/math8k-sft-QwQ-32B-16k-reasoning-traces-limo1500
Viewer
•
Updated
•
1.46k
•
5