chore: add void attention footer links

app.py

@@ -730,13 +730,25 @@ The improvement is largest on asymmetric games where the walkers' complement dis
 [The Void](https://huggingface.co/spaces/forkjoin-ai/the-void) |
 [Buleyean RL](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl) |
 [Glossolalia](https://huggingface.co/spaces/forkjoin-ai/glossolalia) |
+[Glossolalia Examples](https://huggingface.co/spaces/forkjoin-ai/glossolalia-examples) |
 [Metacog](https://huggingface.co/spaces/forkjoin-ai/metacog) |
 [Five Bules](https://huggingface.co/spaces/forkjoin-ai/five-bules) |
 [Quark Personality](https://huggingface.co/spaces/forkjoin-ai/quark-personality)
 
+**Training spaces**:
+[Buleyean RL 70B Trainer](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl-70b-trainer) |
+[Buleyean RL Mistral 7B Trainer](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl-mistral-7b-trainer) |
+[Buleyean RL Qwen 7B Trainer](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl-qwen2.5-7b-trainer) |
+[Buleyean RL DeepSeek R1 7B Trainer](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl-deepseek-r1-7b-trainer) |
+[Buleyean RL Gemma 9B Trainer](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl-gemma2-9b-trainer) |
+[Buleyean RL Qwen 14B Trainer](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl-qwen2.5-14b-trainer) |
+[Buleyean RL Mistral Small 24B Trainer](https://huggingface.co/spaces/forkjoin-ai/buleyean-rl-mistral-small-24b-trainer)
+
 **Source**: [github.com/affectively-ai/aeon](https://github.com/affectively-ai/aeon)
 
 Built by [AFFECTIVELY](https://affectively.ai). The complement distribution was always softmax attention. The void boundary was always the KV cache. We just named the parts.
+
+Copyright 2026 forkjoin.ai
 """