Minor spelling tweaks

Author: DouglasOrr

pages/posts/potm/2026/02/potm.md

@@ -6,12 +6,12 @@ title: 'February Papers: Thinking Depth, Latent Actions, Quantization and Rieman
 merge_potm: true
 ---
 
-The stream of papers never ends, even so, in February our team found 4 we'd like to share
+The stream of papers never ends, even so, in February our team found 4 we'd like to share:
 
-- *Think Deep, Not Just Long: Measuring LLM Reasoning Effort via Deep-Thinking Tokens* investiagtes how many layers are actually needed for each token during autoregressive LM rollouts.
+- *Think Deep, Not Just Long: Measuring LLM Reasoning Effort via Deep-Thinking Tokens* investigates how many layers are actually needed for each token during autoregressive LM rollouts.
 
-- *Factored Latent Action World Models* takes videos that contain multiple objects, and instead of encoding them into one latent state for the whole scene, tries to one latent state per object.
+- *Factored Latent Action World Models* takes videos that contain multiple objects, and instead of encoding them into one latent state for the _whole scene_, employs one latent state _per object_.
 
 - *LATMiX: Learnable Affine Transformations for Microscaling Quantization of LLMs* generalises Hadamard transforms to better handle outliers when block-quantizing LLMs.
 
-- *Riemannian Mean Flow* extends [MeanFlow](https://arxiv.org/pdf/2505.13447) for generating proteins within the coresponding structured spaces, e.g. the space of all residue positions and orientations.
+- *Riemannian Mean Flow* extends [MeanFlow](https://arxiv.org/pdf/2505.13447) for generating proteins within the corresponding structured spaces, e.g. the space of all residue positions and orientations.