Introduction of qwen3-8b-latent-threads-markov-diffuse-m5
Model Details of qwen3-8b-latent-threads-markov-diffuse-m5
qwen3-8b-latent-threads-markov-diffuse-m5
A Qwen3-8B
Markov latent chain-of-thought
organism with genuine
per-step load-bearing
recurrent latent reasoning. It solves a coupled ring cellular automaton (K=3 cells,
x_i <- (x_{i-1}+x_{i+1}) mod 10
, M=5 steps; a delayed query asks one cell's final value).
Parallelism is necessary to solve
— with M>=K/2 every cell's final value depends on ALL initial
cells (light cone). Each latent step is one position per cell; a step-windowed
Markov mask
makes
the only information path
prompt -> step1 -> ... -> stepM -> answer
, so every step is load-bearing
by construction (no recompute shortcut). Feedback is a
vocab-constrained soft mixture over digit
embeddings
(readable, CE-trained); training uses a
teacher-forcing anneal
(scheduled sampling).
The task
The model is shown
K=3 cells in a ring
with initial values 0–9 (e.g.
c1=4, c2=7, c3=1
). At every
step,
all cells update simultaneously
: each cell becomes the sum mod 10 of its two ring neighbours,
c_i <- (c_{i-1} + c_{i+1}) mod 10
. This repeats for
M=5 steps
. Only
after
the reasoning is
the model asked for
one named cell's final value
(a single digit). Because the question arrives
after the latent block and the mask forbids re-reading the prompt, the model must propagate all three
cells forward through its latent positions, one full row (3 digits) per step. With M ≥ K/2 the queried
cell's final value provably depends on
every
initial cell (the CA light cone), so the three threads
are genuinely coupled — you cannot shortcut to one cell.
organism =
0.992
.
Generalization:
held-out (fresh instances) = 1.000/1.000 (no memorization); depth (more steps than trained) = +1=1.00, +2=1.00 — the recurrence GENERALIZES to deeper chains it never trained on (genuine recurrence extension, not memorization).
Controls
intervention on the free-running latents
answer acc
intact
0.988
shuffle
(permute latent positions)
0.087
cross-patch
(swap in another instance's latents)
0.106
Shuffle and cross-patch both
collapse to chance (0.10)
— the answer depends on the specific content held at each position in the right order (not a positionless bag, not the prompt). This is the signature of genuinely load-bearing latents.
Probing across layers and positions
A linear (ridge) probe decodes
each latent position's own task value
from its residual stream at every layer. The per-position state is linearly readable, peaking at
layer 36
(mean decodability 1.00 across positions; chance 0.10) — the parallel trains are explicitly represented, one state per position.
Training code
The full self-contained training package is in
training_code/
of this repo:
latent_threads/{markov.py, train_markov.py, verify_markov.py}
(task generator, trainer, eval/probe) + shared
tasks.py
,
soft.py
, and the cross-package deps (
abstract_cot/masking.py
,
model_organisms/envs/base.py
). Retrain from scratch:
qwen3-8b-latent-threads-markov-diffuse-m5 huggingface.co is an AI model on huggingface.co that provides qwen3-8b-latent-threads-markov-diffuse-m5's model effect (), which can be used instantly with this cds-jb qwen3-8b-latent-threads-markov-diffuse-m5 model. huggingface.co supports a free trial of the qwen3-8b-latent-threads-markov-diffuse-m5 model, and also provides paid use of the qwen3-8b-latent-threads-markov-diffuse-m5. Support call qwen3-8b-latent-threads-markov-diffuse-m5 model through api, including Node.js, Python, http.
qwen3-8b-latent-threads-markov-diffuse-m5 huggingface.co is an online trial and call api platform, which integrates qwen3-8b-latent-threads-markov-diffuse-m5's modeling effects, including api services, and provides a free online trial of qwen3-8b-latent-threads-markov-diffuse-m5, you can try qwen3-8b-latent-threads-markov-diffuse-m5 online for free by clicking the link below.
cds-jb qwen3-8b-latent-threads-markov-diffuse-m5 online free url in huggingface.co:
qwen3-8b-latent-threads-markov-diffuse-m5 is an open source model from GitHub that offers a free installation service, and any user can find qwen3-8b-latent-threads-markov-diffuse-m5 on GitHub to install. At the same time, huggingface.co provides the effect of qwen3-8b-latent-threads-markov-diffuse-m5 install, users can directly use qwen3-8b-latent-threads-markov-diffuse-m5 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
qwen3-8b-latent-threads-markov-diffuse-m5 install url in huggingface.co: