scombat commited on
Commit
a4bfb51
·
verified ·
1 Parent(s): eb5aa2f

Card update

Browse files
Files changed (1) hide show
  1. README.md +3 -1
README.md CHANGED
@@ -18,7 +18,9 @@ tags:
18
  A "System 1" router: it reads a user prompt and returns its complexity score (0.01 to 1.00), its task type (19 classes)
19
  and whether it needs tools. It never picks a model; the deployer maps these outputs to models with rules. The method,
20
  the training data, the full results and the limitations are in the card of the main repository,
21
- [scombat/s1d1-router](https://huggingface.co/scombat/s1d1-router).
 
 
22
 
23
  - Variant: `v4_vocab32k-fp16`, full model with its vocabulary pruned to 32k tokens and a distillation pass.
24
  - Format: PyTorch fp16 (safetensors), 308 MB.
 
18
  A "System 1" router: it reads a user prompt and returns its complexity score (0.01 to 1.00), its task type (19 classes)
19
  and whether it needs tools. It never picks a model; the deployer maps these outputs to models with rules. The method,
20
  the training data, the full results and the limitations are in the card of the main repository,
21
+ [scombat/s1d1-router](https://huggingface.co/scombat/s1d1-router). An open, replayable bench of the variants against three local LLM
22
+ judges on 360 multilingual prompts, with confidence intervals and witnesses, is
23
+ [s1d1-bench](https://github.com/scombat/s1d1-bench).
24
 
25
  - Variant: `v4_vocab32k-fp16`, full model with its vocabulary pruned to 32k tokens and a distillation pass.
26
  - Format: PyTorch fp16 (safetensors), 308 MB.