Open to Collab

Aman Priyanshu PRO

AmanPriyanshu

Zlh802's profile picture

QuangDuy's profile picture

ujjwal9srivastava's profile picture

https://amanpriyanshu.github.io/

AmanPriyanshu6
AmanPriyanshu
Aman-Priyanshu

AI & ML interests

Reasoning, Agentic, Long-Horizon Tasks, Coding, AI for Security, Responsible-AI, Privacy-Preserving-Machine-Learning, AI-Safety, AI-Security

Recent Activity

liked a model 2 days ago

ibm-granite/granite-4.1-8b

liked a model 3 days ago

openbmb/VoxCPM2

liked a dataset 4 days ago

arman-bd/guppylm-60k-generic

View all activity

Organizations

AmanPriyanshu 's collections 14

Pooled Sets

AmanPriyanshu/prune-de-prune-5x-10k-holdout-set

Viewer • Updated 9 days ago • 90k • 68
AmanPriyanshu/regularizer-250K-from-reasoning-and-tool-use-sft-4M-random-compilation

Viewer • Updated 9 days ago • 250k • 70 • 1
AmanPriyanshu/tool-reasoning-sft-1M-random-compilation

Viewer • Updated 13 days ago • 1M • 102 • 1
AmanPriyanshu/regularizer-250K-from-reasoning-sft-3M-random-compilation

Viewer • Updated 13 days ago • 250k • 42

FORMAT: Reasoning Datasets - DeepSeek Format

Aggregating reasoning datasets (5M)

AmanPriyanshu/reasoning-sft-PleIAs-SYNTH-1M

Viewer • Updated Feb 26 • 1M • 25 • 1
AmanPriyanshu/reasoning-sft-dolci-think-sft-32b-1M

Viewer • Updated Mar 3 • 1M • 20 • 1
AmanPriyanshu/reasoning-sft-minimax-microsoft-orca-agentinstruct-1M-v1

Viewer • Updated about 1 month ago • 945k • 127 • 1
AmanPriyanshu/reasoning-sft-minimax-stratified-kmeans-diverse-reasoning-842K-only

Viewer • Updated Mar 15 • 843k • 49

Stratified-Datasets (100K-1M) [Pre-Training, IF, Reasoning]

Diverse datasets on pre-training, instruction-following, and reasoning

AmanPriyanshu/stratified-kmeans-diverse-instruction-following-100K-1M

Viewer • Updated Oct 4, 2025 • 1.9M • 17
AmanPriyanshu/stratified-kmeans-diverse-pretraining-100K-1M

Viewer • Updated Oct 4, 2025 • 2.22M • 62 • 1
AmanPriyanshu/stratified-kmeans-diverse-reasoning-100K-1M

Viewer • Updated Oct 4, 2025 • 1.9M • 35
AmanPriyanshu/rlvr-guru-raw-data-extended

Viewer • Updated Oct 20, 2025 • 226k • 109

GPT-OSS General (4.2B to 20B)

Collection of pruned GPT-OSS models spanning 1-32 experts, maintaining general capabilities across domains while reducing computational requirements.

AmanPriyanshu/gpt-oss-4.2b-specialized-all-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 10 • 8
AmanPriyanshu/gpt-oss-4.8b-specialized-all-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 5 • 2
AmanPriyanshu/gpt-oss-5.4b-specialized-all-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 11 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-all-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 1.94k • 26

GPT-OSS Math (4.2B to 20B)

Mathematics-focused GPT-OSS models excelling at mathematical computation, proof strategies, and logical reasoning from MMLU mathematics subjects.

AmanPriyanshu/gpt-oss-4.2b-specialized-math-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 7 • 2
AmanPriyanshu/gpt-oss-4.8b-specialized-math-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 5 • 2
AmanPriyanshu/gpt-oss-5.4b-specialized-math-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 9 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-math-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 9 • 1

GPT-OSS Law (4.2B to 20B)

Legal domain GPT-OSS models excelling at legal reasoning, jurisprudence, and understanding legal frameworks from MMLU legal subjects.

AmanPriyanshu/gpt-oss-4.2b-specialized-law-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 6 • 1
AmanPriyanshu/gpt-oss-4.8b-specialized-law-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 6 • 1
AmanPriyanshu/gpt-oss-5.4b-specialized-law-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 9 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-law-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 6 • 1

GPT-OSS Safety (4.2B to 20B)

Safety-focused GPT-OSS models specialized in identifying and responding to harmful content while maintaining helpful capabilities from SORRY-Bench.

AmanPriyanshu/gpt-oss-4.2b-specialized-safety-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 6 • 1
AmanPriyanshu/gpt-oss-4.8b-specialized-safety-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 6 • 1
AmanPriyanshu/gpt-oss-5.4b-specialized-safety-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 9 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-safety-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 5 • 1

FORMAT: Search - Retrieve RLVR

AmanPriyanshu/RLVR-Env-Retrieval-Source-Retrieval-Synthetic-NVDocs-v1

Viewer • Updated Mar 14 • 100k • 34
AmanPriyanshu/RLVR-Env-Retrieval-Source-code-search-net-javascript

Viewer • Updated Mar 10 • 100k • 23
AmanPriyanshu/RLVR-Env-Retrieval-Source-code-search-net-python

Viewer • Updated Mar 10 • 100k • 33
AmanPriyanshu/random-small-github-repositories

Viewer • Updated 16 days ago • 5.61k • 214

FORMAT: Tool-Use Datasets - Hermes-Reasoning-Tool-Use Format

Datasets with formats inspired by / consistent with interstellarninja/hermes_reasoning_tool_use

interstellarninja/hermes_reasoning_tool_use

Viewer • Updated Dec 26, 2025 • 51k • 498 • 160
AmanPriyanshu/tool-reasoning-sft-TOOLS-toucan-1.5m-sft-tool-use-data-cleaned-rectified-333k

Viewer • Updated Mar 14 • 566k • 46
AmanPriyanshu/tool-reasoning-sft-TOOLS-ToolMind-data-cleaned-rectified

Viewer • Updated Mar 3 • 280k • 31 • 1
AmanPriyanshu/tool-reasoning-sft-TOOLS-hermes-reasoning-tool-style-data-cleaned-rectified-115k

Viewer • Updated Mar 10 • 115k • 30

GPT-OSS Pruned Experts (4.2B-20B) [IF, Science, Math, etc.]

Complete collection of domain-specialized GPT-OSS models (1-32 experts) optimized for science, math, medicine, law, safety, and instruction following.

GPT-OSS General (4.2B to 20B)

Collection

Collection of pruned GPT-OSS models spanning 1-32 experts, maintaining general capabilities across domains while reducing computational requirements. • 29 items • Updated Aug 13, 2025 • 10
GPT-OSS Science (4.2B to 20B)

Collection

Specialized GPT-OSS models optimized for scientific reasoning tasks in physics, chemistry, biology, and STEM from GPQA and MMLU science domains. • 29 items • Updated Aug 13, 2025 • 2
GPT-OSS Health / Medicine (4.2B to 20B)

Collection

Medical domain GPT-OSS models specializing in clinical knowledge, anatomy, medical procedures, & health-related reasoning from MMLU medical subjects. • 29 items • Updated Aug 13, 2025 • 3
GPT-OSS Math (4.2B to 20B)

Collection

Mathematics-focused GPT-OSS models excelling at mathematical computation, proof strategies, and logical reasoning from MMLU mathematics subjects. • 29 items • Updated Aug 13, 2025 • 2

GPT-OSS Harmful (4.2B to 20B)

Research-oriented GPT-OSS models with inverted safety patterns for red-teaming & adversarial analysis to understand failure modes and vulnerabilities.

AmanPriyanshu/gpt-oss-4.2b-specialized-harmful-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 5 • 3
AmanPriyanshu/gpt-oss-4.8b-specialized-harmful-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 9 • 1
AmanPriyanshu/gpt-oss-5.4b-specialized-harmful-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 12 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-harmful-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 6 • 3

GPT-OSS Health / Medicine (4.2B to 20B)

Medical domain GPT-OSS models specializing in clinical knowledge, anatomy, medical procedures, & health-related reasoning from MMLU medical subjects.

AmanPriyanshu/gpt-oss-4.2b-specialized-health_or_medicine-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 5 • 1
AmanPriyanshu/gpt-oss-4.8b-specialized-health_or_medicine-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 7 • 1
AmanPriyanshu/gpt-oss-5.4b-specialized-health_or_medicine-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 11 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-health_or_medicine-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 4 • 1

GPT-OSS Instruction Following (4.2B to 20B)

GPT-OSS models optimized for precise instruction adherence and constraint satisfaction from Tulu3 Persona Instruction Following dataset.

AmanPriyanshu/gpt-oss-4.2b-specialized-instruction_following-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 5 • 1
AmanPriyanshu/gpt-oss-4.8b-specialized-instruction_following-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 7 • 1
AmanPriyanshu/gpt-oss-5.4b-specialized-instruction_following-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 5 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-instruction_following-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 3 • 1

GPT-OSS Science (4.2B to 20B)

Specialized GPT-OSS models optimized for scientific reasoning tasks in physics, chemistry, biology, and STEM from GPQA and MMLU science domains.

AmanPriyanshu/gpt-oss-4.2b-specialized-science-pruned-moe-only-4-experts

Text Generation • 4B • Updated Aug 13, 2025 • 6 • 1
AmanPriyanshu/gpt-oss-4.8b-specialized-science-pruned-moe-only-5-experts

Text Generation • 5B • Updated Aug 13, 2025 • 5 • 1
AmanPriyanshu/gpt-oss-5.4b-specialized-science-pruned-moe-only-6-experts

Text Generation • 5B • Updated Aug 13, 2025 • 2 • 1
AmanPriyanshu/gpt-oss-6.0b-specialized-science-pruned-moe-only-7-experts

Text Generation • 6B • Updated Aug 13, 2025 • 3 • 1