Update README.md

README.md

@@ -1,63 +1,86 @@
----
-title: README
-emoji: 🔩
-colorFrom: gray
-colorTo: blue
-sdk: static
-pinned: false
----
-
-<div align="center">
-
-# Tinman Lab
-
-### Autonomous Machines. Second-Order Systems.
-
-<sub>AGENT MEMORY · ADVERSARIAL SAFETY · AGENTIC ECONOMY · PERCEPTION SYSTEMS · APPLIED RESEARCH</sub>
-
----
-
-</div>
-
-## Research
-
-### Disposition Distillation — a three-arc negative result
-
-We set out to train *behavioral dispositions* — self-verification, uncertainty
-acknowledgment, feedback integration — into sub-billion-parameter language
-models. Across three independent operator classes (SFT/DPO LoRA imitation,
-attention-head tempering on `o_proj`, and frozen-base hidden-state confidence
-sidecars), no operator moved judge-measured disposition without simultaneously
-damaging content quality or collapsing into stylistic mimicry. The failure is
-consistent across Qwen3-0.6B, Qwen3-1.7B, Qwen3.5-0.8B, Gemma 4 E2B, and
-SmolLM2-1.7B-Instruct.
-
-The contribution is the falsification, a two-failure-mode taxonomy for linear
-hidden-state probes, and a methodological pipeline that converts
-CV-on-same-distribution false positives into honest negatives.
-
-📄 **Paper** — [arXiv:2604.11867](https://arxiv.org/abs/2604.11867)
-🔧 **Artifacts** — [github.com/tinmanlabsl/disposition-distillation](https://github.com/tinmanlabsl/disposition-distillation)
-
-## Open-source proofs of concept
-
-Apache 2.0. Internal product models are not listed here.
-
-### Tinman SmolOmni (MLA)
-
-- [Tinman-SmolOmni-MLA-256M](https://huggingface.co/Tinman-Lab/Tinman-SmolOmni-MLA-256M)
-- [Tinman-SmolOmni-MLA-500M](https://huggingface.co/Tinman-Lab/Tinman-SmolOmni-MLA-500M)
-- [Tinman-SmolOmni-MLA-Toolkit](https://huggingface.co/Tinman-Lab/Tinman-SmolOmni-MLA-Toolkit)
-
-### Tinman Companion (Gemma 4)
-
-- [Tinman-gemma4-companion-merged](https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-merged) — full precision
-- [Tinman-gemma4-companion-gguf](https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-gguf) — GGUF quantized for llama.cpp
-- [Tinman-gemma4-companion-litert-lm](https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-litert-lm) — LiteRT-LM for on-device deployment
-- [Tinman-gemma4-companion-sft](https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-sft) — SFT checkpoint
-- [Tinman-gemma4-companion-dpo](https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-dpo) — DPO checkpoint
-
-## Links
-
-- [Website](https://tinmanlab.com)
-- [GitHub](https://github.com/tinmanlabsl/)
+---
+title: README
+emoji: 🔩
+colorFrom: gray
+colorTo: blue
+sdk: static
+pinned: false
+---
+<div class="center">
+  <h1>Tinman Lab</h1>
+  <h3>Autonomous Machines. Second-Order Systems.</h3>
+  <div class="sub">AGENT MEMORY · ADVERSARIAL SAFETY · AGENTIC ECONOMY · PERCEPTION SYSTEMS · APPLIED RESEARCH</div>
+</div>
+
+<hr>
+
+<p>We build on-device AI systems that reason, remember, and self-correct — small models designed to run autonomously at the edge with calibrated uncertainty and adversarial robustness.</p>
+
+<h2>Research Areas</h2>
+
+<ul>
+  <li><strong>Agent Memory</strong> — Encrypted semantic memory infrastructure for persistent agent context</li>
+  <li><strong>Adversarial Safety</strong> — Multi-agent stress-testing and trust verification for autonomous systems</li>
+  <li><strong>Perception Systems</strong> — On-device vision, voice, and multimodal understanding</li>
+  <li><strong>Disposition Distillation</strong> — A three-arc study finding that imitation, attention-head tempering, and frozen-base sidecars all fail to move judge-measured disposition without damaging content quality at sub-billion scale (<a href="https://arxiv.org/abs/2604.11867">arXiv:2604.11867</a>).</li>
+</ul>
+
+<h2>Open-Source Releases</h2>
+
+<p><strong>Tinman SmolOmni (MLA)</strong> — small omnimodal models with multi-head latent attention.</p>
+
+<table>
+  <thead>
+    <tr><th>Model</th><th>Description</th></tr>
+  </thead>
+  <tbody>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-SmolOmni-MLA-256M">Tinman-SmolOmni-MLA-256M</a></td><td>256M parameter omnimodal</td></tr>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-SmolOmni-MLA-500M">Tinman-SmolOmni-MLA-500M</a></td><td>500M parameter omnimodal</td></tr>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-SmolOmni-MLA-Toolkit">Tinman-SmolOmni-MLA-Toolkit</a></td><td>Training and inference toolkit</td></tr>
+  </tbody>
+</table>
+
+<p><strong>Tinman Companion</strong> — Gemma 4 fine-tunes for on-device companion use cases.</p>
+
+<table>
+  <thead>
+    <tr><th>Model</th><th>Description</th></tr>
+  </thead>
+  <tbody>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-merged">Tinman-gemma4-companion-merged</a></td><td>Full-precision merged model</td></tr>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-gguf">Tinman-gemma4-companion-gguf</a></td><td>GGUF quantized for llama.cpp</td></tr>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-litert-lm">Tinman-gemma4-companion-litert-lm</a></td><td>LiteRT-LM for on-device deployment</td></tr>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-sft">Tinman-gemma4-companion-sft</a></td><td>SFT checkpoint</td></tr>
+    <tr><td><a href="https://huggingface.co/Tinman-Lab/Tinman-gemma4-companion-dpo">Tinman-gemma4-companion-dpo</a></td><td>DPO checkpoint</td></tr>
+  </tbody>
+</table>
+
+<h2>Links</h2>
+
+<p class="links">
+  <a href="https://tinmanlab.com" rel="nofollow">Website</a>
+  <a href="https://github.com/tinmanlabsl/" rel="nofollow">GitHub</a>
+  <a href="https://arxiv.org/abs/2604.11867" rel="nofollow">Paper (arXiv)</a>
+  <a href="https://github.com/tinmanlabsl/disposition-distillation" rel="nofollow">DD Artifacts</a>
+</p>
+
+<style>
+  body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, 'Helvetica Neue', Arial, sans-serif; max-width: 720px; margin: 0 auto; padding: 2rem 1.5rem; color: #1a1a2e; line-height: 1.6; background: #fff; }
+  .center { text-align: center; }
+  h1 { font-size: 2rem; margin-bottom: 0.2rem; font-weight: 700; }
+  h3 { font-size: 1.1rem; font-weight: 400; color: #333; margin-top: 0.2rem; margin-bottom: 0.3rem; }
+  .sub { font-size: 0.75rem; letter-spacing: 0.15em; color: #888; margin-bottom: 1.5rem; }
+  hr { border: none; border-top: 1px solid #e0e0e0; margin: 1.5rem 0; }
+  h2 { font-size: 1.2rem; margin-top: 1.8rem; margin-bottom: 0.8rem; }
+  p { margin-bottom: 0.8rem; }
+  ul { padding-left: 1.2rem; }
+  li { margin-bottom: 0.4rem; }
+  table { width: 100%; border-collapse: collapse; margin: 1rem 0; font-size: 0.95rem; }
+  th, td { text-align: left; padding: 0.5rem 0.8rem; border-bottom: 1px solid #eee; }
+  th { font-weight: 600; border-bottom: 2px solid #ddd; }
+  a { color: #2563eb; text-decoration: none; }
+  a:hover { text-decoration: underline; }
+  .links a { margin-right: 1.5rem; }
+  strong { font-weight: 600; }
+</style>
+