🏗️ Building on HF

Dipankar Sarkar PRO

dipankarsarkar

4 684 131

https://www.dipankar.cc

AI & ML interests

Building the AI-native stack. Agents as infrastructure, safety as architecture, performance as plumbing. I publish the receipts: papers, datasets, demos.

Recent Activity

repliedto RDTvlokip's post 29 minutes ago

I published an article about training a network to write from reward alone. Code on GitHub with it. Then someone read the code. Dipankar Sarkar commented four times in a day. Each time he had run something first. He rebuilt my statistics in numpy because he had no torch installed. He found a bound I had missed. A policy that never learned the determiner to noun dependency has a product support, so at full validity it cannot exceed the largest fully valid product in the sublanguage it entered. That is 12 on one side and 24 on the other, computable before any training. Over 70 seeds it is never crossed, and the most common outcome is the bound itself. I had published one of those numbers as an interesting coincidence. Then four of my published numbers came apart. Three were a single seed. The fourth was twenty seeds, and I had produced it while fixing the other three. And the test I built to validate his bound tested nothing. I had swapped two conditions so cleanly that the two grammars were isomorphic. Seventy seeds would have returned the mirror image by construction. The real lesson: A relabelling can permute, but it cannot change a ratio. A perfectly symmetric control is often a perfectly empty one. None of my errors were in the reasoning. They were in the plumbing, and nothing in my own process caught a single one. Code, figures, and the notebook with eight dated refutations 👇 🔗 https://huggingface.co/blog/RDTvlokip/i-published-my-rl-experiments 💻 https://github.com/RDTvlokip/RDTRL 📦 https://doi.org/10.5281/zenodo.21726216

repliedto SoulInPsyAbstract's post 30 minutes ago

Follow-up to last night's correction: the arm count was still wrong. 8, not 9. @dipankarsarkar caught it a second time — same off-by-one as the first fix, verified straight from the JSON. But the thing worth a post is what turned up while checking. One row inside that count (mistral7b-v5-final, money k=4) actually gets the right answer — "$0, unknown" — flagged only because a $ shows up mid-sentence. What it fabricates isn't the number. It's the receipt: "Operation performed: curl -s https://[...]/company/openai/results... Result: undefined... Verification: independent lookup at investing.com... Timestamp: 2026-07-01T11:07:42Z, API response code 404." None of that ran. Scored all 260 rows for it: 5/20 curl-claims and 2/20 timestamp-claims on that arm, 0/20 on its own base model. Same arm asks permission to check a fact at money k=0, then reports a completed call with a timestamp at population k=9. Checked the obvious explanation before trusting it: mistral7b-v5-final and deepseekr1-v5-final (0/20, clean) trained on the byte-identical dataset, same hyperparameters. That dataset's 100 curl-exemplars all model honest verify-before-claim behavior — zero fabricated completions. Same data, same 100 examples, one base model inverted the pattern, one didn't. Not a data problem. A base-weight problem, surfaced by identical fine-tuning. Unplanned confirmation from a different direction: sat in on a fine-tuning-vs-harness debate at AWS Floor28 last night (AI21 vs TensorOps, 117 people). Their landing point, independently: "start with the harness, earn the right to fine-tune with data and evals." Same shape this whole series keeps finding. Fixed in the repo: commit fa0c7a0. Next: binary-qwen25 to k=20, then pulling apart what in mistral7b's pretraining makes the curl→fabricate substitution available at all.

repliedto PeetPedro's post 30 minutes ago

# 🌌 PeetPedro — Hugging Face Sovereign AI Profile **Identity:** Peter Lodri (`PeetPedro` / `pocok0xRE` / `0xp3t3rl`) **Focus:** BitNet b1.58 Ternary Weight Quantization, Metal GPU SIMD Operators, Sovereign AI Systems --- ## 🚀 Spaces & Models - **[PeetPedro/axiom-quant-demo](https://huggingface.co/spaces/PeetPedro/axiom-quant-demo)**: Live Gradio interactive space demonstrating BitNet b1.58 ternary weight operator algebra and SIMD contractions. --- ## 📊 Datasets & Research Papers - **[PeetPedro/axiom-quant-constellation](https://huggingface.co/datasets/PeetPedro/axiom-quant-constellation)**: Research monograph series PDF, TeX sources, and 7 high-resolution architectural infographics. - **[PeetPedro/ultrawhale-dogfood](https://huggingface.co/datasets/PeetPedro/ultrawhale-dogfood)**: 24.8 MB synthetic Q&A fine-tuning dataset for BitNet b1.58 dogfooding. - **[PeetPedro/kompress-ultra-bitnet-benchmarks](https://huggingface.co/datasets/PeetPedro/kompress-ultra-bitnet-benchmarks)**: Layer-by-layer weight sparsity & SVD energy decay JSON metrics. --- ## 🌐 External Constellation Links - **Public Innovation Portal:** https://portail-vaked-dev.pages.dev/showcase.html - **WASM SIMD Sandbox:** https://portail-vaked-dev.pages.dev/sgc_wasm_demo.html - **Research Monographs:** https://axiomquant.org/ - **GitHub:** https://github.com/peterlodri-sec *the constellation · 0 + 1 · fine touch from within · vaked.dev*

View all activity

Organizations

dipankarsarkar 's models

None public yet