Upload README.md with huggingface_hub

README.md

@@ -4,11 +4,11 @@ language:
 license: apache-2.0
 base_model: Qwen/Qwen3-1.7B
 tags:
+- qthink
 - continuous-thought
 - latent-reasoning
 - distillation
 - gsm8k
-- codi
 datasets:
 - openai/gsm8k
 metrics:
@@ -40,13 +40,15 @@ model-index:
 
 ## Overview
 
-This model implements **Continuous Thought Distillation (CODI)** — an autoregressive latent
-reasoning loop that processes K=6 continuous thought steps before generating a text answer.
-Teacher hidden states are extracted from multiple chain-of-thought rollouts generated by the
-base model and distilled into the latent representations.
+This model implements **QThink** (Parallel Latent Reasoning via Per-Step Distillation of
+Multiple Rollouts) — an autoregressive latent reasoning loop that processes K=6 continuous
+thought steps before generating a text answer. Teacher hidden states are extracted from
+multiple chain-of-thought rollouts generated by the base model and distilled into the latent
+representations at every step.
 
 **Key idea**: Instead of distilling from a single reasoning trace, we aggregate hidden states
-from multiple rollouts (16 per problem). The teacher signal is the uniform average of ALL rollouts (correct + incorrect).
+from multiple rollouts (16 per problem) and supervise every latent step. The teacher signal
+is the uniform average of ALL rollouts (correct + incorrect).
 
 ## Architecture
 
@@ -178,9 +180,9 @@ chain-of-thought text after the latent steps, requiring more tokens than standar
 | **This model** | **codi_uniform** | **True** | **2.0** | **256** | **83.2%** |
 | Qwen3-1.7B (base) | — | — | — | — | 77.3% |
 | Discrete SFT | sft | — | — | — | 80.7% |
-| CODI RW final-step | rw | no | 1.0 | 128 | 81.0% |
-| CODI Uniform per-step ans256 | uniform | yes | 2.0 | 256 | 83.2% |
-| CODI RW per-step ans256 | rw | yes | 1.0 | 256 | 82.7% |
+| QThink RW final-step | rw | no | 1.0 | 128 | 81.0% |
+| QThink Uniform per-step ans256 | uniform | yes | 2.0 | 256 | 83.2% |
+| QThink RW per-step ans256 | rw | yes | 1.0 | 256 | 82.7% |
 
 ## Citation
 
@@ -188,7 +190,7 @@ If you use this model, please cite:
 
 ```bibtex
 @misc{continuous-thought-2025,
-  title={Continuous Thought Distillation: Reward-Weighted Multi-Trace Reasoning},
+  title={QThink: Parallel Latent Reasoning via Per-Step Distillation of Multiple Rollouts},
   author={Lakshya Agrawal},
   year={2025},
   url={https://huggingface.co/LakshyAAAgrawal/continuous-thought-r11_uniform_perstep_g2_ans256}