Qwen3.5-Embedding-0.8B

An embedding model fine-tuned from Qwen/Qwen3.5-0.8B-Base for OpenClaw memory retrieval.

The repository name follows the official Qwen/Qwen3-Embedding-0.6B naming pattern while preserving the Qwen3.5 base-model version. The name does not encode a training-stage label. The uploaded weights are the final 5-epoch fine-tuned result from this project.

Intended use

Chinese/English semantic retrieval for OpenClaw-style memory search, including short keyword queries and natural questions.

Encoding protocol

Query:

Instruct: Given a query, retrieve relevant passages that answer the query
Query: {query}

Memory/passages: no instruction prefix.

Use last-token pooling and L2 normalization. The model supports Matryoshka-style dimensions tested at 128, 256, 512, 768, and 1024. This is not a sentence-transformers package with a built-in pooling wrapper; implement the pooling and query prefix as described above.

Comparison with Qwen/Qwen3-Embedding-0.6B

All values are Recall@K or MRR. S2 is this model; Official is Qwen/Qwen3-Embedding-0.6B. The same tokenizer protocol, query instruction, pooling, normalization, corpus, and query set were used for each row.

OpenClaw fine-tuning validation

Dim	S2 R@1	Official R@1	S2 R@3	Official R@3	S2 R@5	Official R@5	S2 R@10	Official R@10	S2 MRR	Official MRR
128	8.96%	17.19%	19.57%	31.84%	25.82%	37.98%	35.32%	46.49%	17.57%	27.38%
256	15.04%	20.12%	30.74%	35.99%	38.58%	43.67%	50.14%	52.35%	26.46%	31.25%
512	21.67%	22.33%	40.57%	39.19%	49.75%	47.21%	62.02%	57.16%	34.84%	34.11%
768	23.88%	22.94%	44.67%	40.57%	53.45%	48.31%	65.06%	58.71%	37.60%	34.97%
1024	25.21%	23.66%	45.55%	41.46%	54.62%	49.36%	66.56%	59.92%	38.86%	35.66%

Unseen OpenClaw memory

The corpus contains 382 chunks from memory dated after 2026-08-08; the benchmark contains 1,048 Gemini-generated queries with multi-positive qrels where applicable.

Dim	S2 R@1	Official R@1	S2 R@3	Official R@3	S2 R@5	Official R@5	S2 R@10	Official R@10	S2 MRR	Official MRR
128	12.02%	18.80%	26.15%	40.55%	35.50%	49.43%	47.81%	59.73%	23.76%	33.14%
256	17.08%	20.32%	35.40%	43.89%	45.99%	51.91%	61.83%	62.02%	31.00%	35.07%
512	20.42%	23.09%	43.23%	45.32%	54.87%	54.10%	68.61%	63.93%	36.48%	37.40%
768	21.56%	22.04%	45.80%	45.80%	59.06%	54.39%	72.23%	63.84%	38.22%	36.91%
1024	22.33%	21.09%	46.47%	45.52%	58.59%	53.91%	72.04%	64.31%	38.80%	36.25%

Conventional content validation sample

A stratified sample of 382 queries from the pretraining validation split, with 761 unique positive/negative passages.

Dim	S2 R@1	Official R@1	S2 R@3	Official R@3	S2 R@5	Official R@5	S2 R@10	Official R@10	S2 MRR	Official MRR
128	80.37%	81.41%	93.46%	93.19%	96.60%	96.34%	98.43%	97.91%	87.45%	87.93%
256	87.70%	85.86%	96.60%	93.98%	98.43%	96.34%	99.74%	97.64%	92.42%	90.43%
512	89.53%	85.34%	97.91%	93.72%	99.21%	96.34%	99.74%	98.43%	93.71%	90.22%
768	91.10%	84.82%	97.64%	95.03%	99.21%	97.38%	99.74%	98.69%	94.58%	90.34%
1024	91.36%	84.29%	97.64%	95.29%	98.95%	97.12%	99.74%	98.69%	94.74%	90.13%

Recommendations

• For OpenClaw's default memory_search behavior, the effective default is up to 10 results; use Recall@10 as the primary metric.
• Use 1024 dimensions for maximum ranking quality, especially when top-1/top-3 and MRR matter.
• Use 768 dimensions as the storage/Recall@5-10 compromise; it uses about 25% less vector storage than 1024 dimensions.

Base model and license

This model is derived from Qwen/Qwen3.5-0.8B-Base. Please review and comply with the base model's license and usage requirements. The license field is intentionally omitted here until the base model license is confirmed from its authoritative model card.