Generation Presets
Saved sampling policy. When a preset is enabled and assigned, the proxy enforces its defined fields upstream. Presets are never assigned to model profiles automatically.
When speculative-decoding safe is enabled on a preset, the proxy explicitly sends
min_p=0 and removes logit_bias immediately before forwarding upstream. Preset assignment remains entirely manual.Preset Inventory
| Preset | Default | Enabled | Spec-safe | Sampling | Notes |
|---|---|---|---|---|---|
Deterministic Codingdeterministic_coding | False | True | False | temperature=0.15 top_p=0.9 top_k=40 min_p=0.03 repetition_penalty=1.05 seed=null | Low-variance coding and patching. |
Balanced Codingbalanced_coding | False | True | False | temperature=0.65 top_p=0.9 top_k=50 min_p=0.05 repetition_penalty=1.05 seed=null | Default profile for Zoo/local coding: varied but controlled. |
General Assistantgeneral_assistant | False | True | False | temperature=0.75 top_p=0.92 top_k=60 min_p=0.03 repetition_penalty=1.04 seed=null | General chat/default assistant work. |
Creative Brainstormingcreative_brainstorming | False | True | False | temperature=1 top_p=0.95 top_k=100 min_p=0.02 repetition_penalty=1.02 seed=null | Higher variance idea generation. |
Tool Call Safetool_call_safe | False | True | False | temperature=0.2 top_p=0.85 top_k=40 min_p=0.03 repetition_penalty=1.05 seed=null | Conservative output for tool-call-heavy work. |
DeepSeek Speculativedeepseek_speculative | False | True | False | temperature=0.65 top_p=0.9 top_k=50 min_p= repetition_penalty=1.05 seed=null | DeepSeek V4 DSpark speculative decoding; min_p omitted. |
GLM 5.3 Flash Speculativeglm_5_3_flash_speculative | True | True | True | temperature=1 top_p=0.95 top_k= min_p=0 repetition_penalty= seed=null | GLM 5.3 Flash speculative/MTP compatibility preset. Uses GLM temperature=1.0 and top_p=0.95 defaults; forces min_p=0 and strips logit_bias before upstream forwarding. |