W
wiiiktor/chopper_demo
70-second video: youtu.be/KMvmdXCEOJU
★ 5·updated 2026-11-18
View on GitHub ↗Direct link · no redirect page
We have neither an install command nor a public remote endpoint from this repository, so we do not print a config. A made-up one is worse than none. How each client wires MCP up →
- Star count
- 5 ★ · wiiiktor/chopper_demo · read from the GitHub search API · (checked 2026-11-19 20:00:48)
- Source
- collected from the punkpeye/awesome-mcp-servers list · type: MCP server
- Harvest history
- first seen 2026-11-18 · last seen 2026-11-20 · seen 3× · listed by 3 sources
- Score
- score 300 · 5 ★ · 3 sources · mcp
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: MCP servers. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
wiiiktor/chopper_demo;first catalogued 2026-11-18;last seen 2026-11-20。There is no official certification or quality rating here — we record sources and dates only.README excerpt (English, untranslated) · Read the full README on GitHub →
70-second video: youtu.be/KMvmdXCEOJU
Excerpted from the repository README (abridged)
Try it in your browser, nothing to install: English demo · Polish demo They run the recommended base chopper (0.31B) on your CPU, with small (0.14B) as a faster option.
See it on real examples: 20 English questions, Chopper vs Opus 5.5 vs gold (PDF)
> Where this is going. The standalone chopper is a first, practical step. The main idea is an integrated chopper: chopping built into the answering LLM itself, reading the model's own cache of the session instead of re-encoding the text in a second model, which should give the highest possible quality and speed. A proof of concept with MiniCPM3-4B is described here.
## Chopper 0.3B close to Opus 5.5 in cutting out the best pieces
Chops are pieces cut out of the original text snippets retrieved from a database or a long conversation: exact quotes,
never rewritten. The chopper (mmBERT, 0.31B) makes them; for comparison, Claude Opus 5.5 makes them too, also forced to
quote directly.
> What is new: a new method, invented in November 2026 in Poland (EU). The chopper is not an LLM summarising the snippets:
> a small encoder reads the question and all 20 snippets together in one pass and marks every word keep or drop. There is no
> text generation, so it is fast: 50 ms per question (20 snippets, ~2,400 tokens) on a consumer-grade GPU (RTX 3090), 36 ms with mmBERT-small.
> One database retrieval per question, even at 4 hops. Every result below uses a single retrieval: each question gets its
> 20 snippets once, the chopper cuts them once, and Qwen3-14B answers once. There are no follow-up searches and no second
> rounds, so a 4-hop question has to be chained from one cut.
### A reranker that also gives you the chops
One model replaces both the reranker and the compressor. Full comparison: below.
### Qwen3-14B answering from full snippets or just short chops
400 MuSiQue questions, 100 each at 1, 2, 3 and 4 hops, each with its own 20 retrieved snippets. Qwen3-14B (4-bit, no fine-tuning,
no thinking mode) answers from the given text only; gpt-5.6-luna judges. Length = prompt tokens. Each score counts every correct
answer, including those from incomplete chops; the number in brackets counts only questions whose chops kept every hop and the answer.
### Qwen3-14B: all snippets vs chops from the chopper vs chops from Opus 5.5
Read each diagram like website traffic: questions flow left to right, and at every hop the ones whose fact the chops missed drop
out; the last stage is Qwen3-14B answering correctly. Top: all 20 retrieved snippets. Middle: the chopper's chops. Bottom: Opus 5.5's chops.
The last stage is the table's score; the lighter block under it counts answers Qwen got right from incomplete chops, and the
number in brackets leaves them out. More: docs/DETAILS.md §7.
### Also a reranker: ranking and chopping in one pass
The chopper scores every passage while it chops (mean p(KEEP) of the passage's 16 most likely tokens). On the
8-language rerank holdout (400 queries, 16 passages each, never trained on), ranked against dedicated rerankers:
| ranking 16 passages | nDCG@10 | Hit@1 | GPU time per query (RTX 3090) |
| chopper, mmBERT-base (0.31B), one pass | 0.973 | 0.972 | 0.20 s, chops + ANSWER/SEARCH included |
See it on real examples: 20 English questions, Chopper vs Opus 5.5 vs gold (PDF)
> Where this is going. The standalone chopper is a first, practical step. The main idea is an integrated chopper: chopping built into the answering LLM itself, reading the model's own cache of the session instead of re-encoding the text in a second model, which should give the highest possible quality and speed. A proof of concept with MiniCPM3-4B is described here.
## Chopper 0.3B close to Opus 5.5 in cutting out the best pieces
Chops are pieces cut out of the original text snippets retrieved from a database or a long conversation: exact quotes,
never rewritten. The chopper (mmBERT, 0.31B) makes them; for comparison, Claude Opus 5.5 makes them too, also forced to
quote directly.
> What is new: a new method, invented in November 2026 in Poland (EU). The chopper is not an LLM summarising the snippets:
> a small encoder reads the question and all 20 snippets together in one pass and marks every word keep or drop. There is no
> text generation, so it is fast: 50 ms per question (20 snippets, ~2,400 tokens) on a consumer-grade GPU (RTX 3090), 36 ms with mmBERT-small.
> One database retrieval per question, even at 4 hops. Every result below uses a single retrieval: each question gets its
> 20 snippets once, the chopper cuts them once, and Qwen3-14B answers once. There are no follow-up searches and no second
> rounds, so a 4-hop question has to be chained from one cut.
### A reranker that also gives you the chops
One model replaces both the reranker and the compressor. Full comparison: below.
### Qwen3-14B answering from full snippets or just short chops
400 MuSiQue questions, 100 each at 1, 2, 3 and 4 hops, each with its own 20 retrieved snippets. Qwen3-14B (4-bit, no fine-tuning,
no thinking mode) answers from the given text only; gpt-5.6-luna judges. Length = prompt tokens. Each score counts every correct
answer, including those from incomplete chops; the number in brackets counts only questions whose chops kept every hop and the answer.
### Qwen3-14B: all snippets vs chops from the chopper vs chops from Opus 5.5
Read each diagram like website traffic: questions flow left to right, and at every hop the ones whose fact the chops missed drop
out; the last stage is Qwen3-14B answering correctly. Top: all 20 retrieved snippets. Middle: the chopper's chops. Bottom: Opus 5.5's chops.
The last stage is the table's score; the lighter block under it counts answers Qwen got right from incomplete chops, and the
number in brackets leaves them out. More: docs/DETAILS.md §7.
### Also a reranker: ranking and chopping in one pass
The chopper scores every passage while it chops (mean p(KEEP) of the passage's 16 most likely tokens). On the
8-language rerank holdout (400 queries, 16 passages each, never trained on), ranked against dedicated rerankers:
| ranking 16 passages | nDCG@10 | Hit@1 | GPU time per query (RTX 3090) |
| chopper, mmBERT-base (0.31B), one pass | 0.973 | 0.972 | 0.20 s, chops + ANSWER/SEARCH included |
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/wiiiktor-chopper-demo/)
- Source
- awesome · https://github.com/wiiiktor/chopper_demo
- Browse by field
- MCP servers
- Traction
- 5 ★
- Listed by
- 3 independent sources
- added
- today
More MCP servers
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.