Instructions to use yava-code/Tessera-135M-Gate with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use yava-code/Tessera-135M-Gate with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="yava-code/Tessera-135M-Gate", trust_remote_code=True)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("yava-code/Tessera-135M-Gate", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use yava-code/Tessera-135M-Gate with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "yava-code/Tessera-135M-Gate"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yava-code/Tessera-135M-Gate",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/yava-code/Tessera-135M-Gate

SGLang

How to use yava-code/Tessera-135M-Gate with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "yava-code/Tessera-135M-Gate" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yava-code/Tessera-135M-Gate",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "yava-code/Tessera-135M-Gate" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yava-code/Tessera-135M-Gate",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use yava-code/Tessera-135M-Gate with Docker Model Runner:
```
docker model run hf.co/yava-code/Tessera-135M-Gate
```

Tessera-135M-Gate / README.md

yava-code

Rewrite card intro: lead with the story, move numbers to a section

4c542e3 verified 19 days ago

preview code

Raw History Blame Contribute Delete

3.77 kB

---
library_name: transformers
pipeline_tag: text-generation
base_model: HuggingFaceTB/SmolLM2-135M
tags:
- next-concept-prediction
- conceptlm
- causal-lm
- smollm2
- tessera
license: apache-2.0
---
# Tessera-135M-Gate

The smallest member of the Tessera family, and the reason the bigger ones exist.
This 135M `architecture-gate checkpoint` (a `HuggingFaceTB/SmolLM2-135M` backbone
with the full Next Concept Prediction path) was trained on a deliberately overfit
TinyStories subset before any large spend, to prove the whole path - concept
pooling, quantized codes, causal concept blocks, feedback - trains end to end.
It is not a language-model quality result, and it is not meant to be: what makes
it worth publishing is what broke, and what the break predicted at scale.

## TL;DR

Small and cheap on purpose: this gate run exists to de-risk the 1B-token comparison before spending on it. It earned its keep in both directions.

- **All three objectives train** (NTP, NCP, VQ fell 41 to 43% over the run),
  and the decoder causally uses the concept channel even at 135M: zeroing the
  feedback costs +0.0205 nats of held-out NTP loss.
- **It also caught a real failure mode**: effective codebook perplexity stayed
  near 2.4 with usage around a third of the codebook, and shuffled feedback  cost nothing - a low-entropy shortcut the 1B-token run later left far behind.  Gate small, then scale: some behaviors only appear above a scale threshold.


## Numbers

The raw record this narrative is built from; per-arm READMEs and the whitepaper hold the
full analysis.

| Metric | Value |
| --- | ---: |
| Held-out NTP loss | 1.8751 |
| Held-out perplexity | 6.5213 |
| Zero feedback delta | 0.0205 |
| Shuffled feedback delta | 0.0000 |
| Codebook perplexity / usage | 2.4087 / 33.0% |
| Training tokens | 16,777,216 |
| Tracked compute estimate | $1.38 |

Intervention deltas are increases in held-out NTP loss relative to normal predicted concept
feedback, evaluated on identical batches.

## Architecture

- chunk size: 4
- product code: 9 segments x 64 entries
- causal concept blocks: 2
- injection point: before token decoder block 2
- NCP target: next continuous concept
- loss: `L_ntp + 1 L_ncp + 1 L_vq`

This is a compact ConceptLM-style implementation, not an 8.9B NCP-ArchPreview replica. It
omits iterative residual coding, cross-scale residual connections, and the large-scale
training recipe.

## Training data and provenance

The matched corpus and the training record are pinned in the ncp-smol repository:
packed-cache SHA256 hashes, the complete metric log, trainer state, and the raw
intervention evaluation JSON (also shipped in this repository as `eval.json` and
`metrics.jsonl`).

## Loading

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("yava-code/Tessera-135M-Gate")
model = AutoModelForCausalLM.from_pretrained("yava-code/Tessera-135M-Gate", trust_remote_code=True)
```

## The Tessera family

Three checkpoints, one story, in reading order:

- [Tessera-135M-Gate](https://huggingface.co/yava-code/Tessera-135M-Gate) - 135M architecture gate (overfit, published for the scale story) **(this model)**
- [Tessera-1B-Nano-Base](https://huggingface.co/yava-code/Tessera-1B-Nano-Base) - 1B-token matched NTP-only control
- [Tessera-1B-Nano](https://huggingface.co/yava-code/Tessera-1B-Nano) - 1B-token matched concept arm

All cards are generated from the run artifacts by the same `build_card`; the
[study repository](https://github.com/yava-code/Tessera-1B-Nano) holds the
whitepaper and full records.
## References

- Code and study: https://github.com/yava-code/Tessera-1B-Nano
- ConceptLM: https://arxiv.org/abs/2602.08984
- NCP-ArchPreview: https://arxiv.org/abs/2609.10715

	---
	library_name: transformers
	pipeline_tag: text-generation
	base_model: HuggingFaceTB/SmolLM2-135M
	tags:
	- next-concept-prediction
	- conceptlm
	- causal-lm
	- smollm2
	- tessera
	license: apache-2.0
	---

	# Tessera-135M-Gate

	The smallest member of the Tessera family, and the reason the bigger ones exist.
	This 135M `architecture-gate checkpoint` (a `HuggingFaceTB/SmolLM2-135M` backbone
	with the full Next Concept Prediction path) was trained on a deliberately overfit
	TinyStories subset before any large spend, to prove the whole path - concept
	pooling, quantized codes, causal concept blocks, feedback - trains end to end.
	It is not a language-model quality result, and it is not meant to be: what makes
	it worth publishing is what broke, and what the break predicted at scale.

	## TL;DR

	Small and cheap on purpose: this gate run exists to de-risk the 1B-token comparison before spending on it. It earned its keep in both directions.

	- All three objectives train (NTP, NCP, VQ fell 41 to 43% over the run),
	and the decoder causally uses the concept channel even at 135M: zeroing the
	feedback costs +0.0205 nats of held-out NTP loss.
	- It also caught a real failure mode: effective codebook perplexity stayed
	near 2.4 with usage around a third of the codebook, and shuffled feedback cost nothing - a low-entropy shortcut the 1B-token run later left far behind. Gate small, then scale: some behaviors only appear above a scale threshold.


	## Numbers

	The raw record this narrative is built from; per-arm READMEs and the whitepaper hold the
	full analysis.

	\| Metric \| Value \|
	\| --- \| ---: \|
	\| Held-out NTP loss \| 1.8751 \|
	\| Held-out perplexity \| 6.5213 \|
	\| Zero feedback delta \| 0.0205 \|
	\| Shuffled feedback delta \| 0.0000 \|
	\| Codebook perplexity / usage \| 2.4087 / 33.0% \|
	\| Training tokens \| 16,777,216 \|
	\| Tracked compute estimate \| $1.38 \|

	Intervention deltas are increases in held-out NTP loss relative to normal predicted concept
	feedback, evaluated on identical batches.

	## Architecture

	- chunk size: 4
	- product code: 9 segments x 64 entries
	- causal concept blocks: 2
	- injection point: before token decoder block 2
	- NCP target: next continuous concept
	- loss: `L_ntp + 1 L_ncp + 1 L_vq`

	This is a compact ConceptLM-style implementation, not an 8.9B NCP-ArchPreview replica. It
	omits iterative residual coding, cross-scale residual connections, and the large-scale
	training recipe.

	## Training data and provenance

	The matched corpus and the training record are pinned in the ncp-smol repository:
	packed-cache SHA256 hashes, the complete metric log, trainer state, and the raw
	intervention evaluation JSON (also shipped in this repository as `eval.json` and
	`metrics.jsonl`).

	## Loading

	```python
	from transformers import AutoModelForCausalLM, AutoTokenizer

	tokenizer = AutoTokenizer.from_pretrained("yava-code/Tessera-135M-Gate")
	model = AutoModelForCausalLM.from_pretrained("yava-code/Tessera-135M-Gate", trust_remote_code=True)
	```

	## The Tessera family

	Three checkpoints, one story, in reading order:

	- [Tessera-135M-Gate](https://huggingface.co/yava-code/Tessera-135M-Gate) - 135M architecture gate (overfit, published for the scale story) (this model)
	- [Tessera-1B-Nano-Base](https://huggingface.co/yava-code/Tessera-1B-Nano-Base) - 1B-token matched NTP-only control
	- [Tessera-1B-Nano](https://huggingface.co/yava-code/Tessera-1B-Nano) - 1B-token matched concept arm

	All cards are generated from the run artifacts by the same `build_card`; the
	[study repository](https://github.com/yava-code/Tessera-1B-Nano) holds the
	whitepaper and full records.
	## References

	- Code and study: https://github.com/yava-code/Tessera-1B-Nano
	- ConceptLM: https://arxiv.org/abs/2602.08984
	- NCP-ArchPreview: https://arxiv.org/abs/2609.10715