From 7678c7d5e84b8f46ad9a426f2b0386f4e672e6f1 Mon Sep 17 00:00:00 2001
From: Gahow Wang <gahow.wang@gmail.com>
Date: Mon, 15 Jun 2026 20:35:23 +0800
Subject: [PATCH] Switch 27B TP A/B to length-aware TTFT SLO (4s + L_in/8k),
 widen search

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---
 configs/examples/dash0_qwen27b_tp_ab.json | 19 ++++---------------
 1 file changed, 4 insertions(+), 15 deletions(-)

diff --git a/configs/examples/dash0_qwen27b_tp_ab.json b/configs/examples/dash0_qwen27b_tp_ab.json
index 9cb15f0..ee2ceda 100644
--- a/configs/examples/dash0_qwen27b_tp_ab.json
+++ b/configs/examples/dash0_qwen27b_tp_ab.json
@@ -145,20 +145,9 @@
   "slo": {
     "target_pass_rate": 0.95,
     "ttft_rule": {
-      "kind": "step_ms",
-      "buckets": [
-        {
-          "max_input_tokens": 4096,
-          "threshold_ms": 2000
-        },
-        {
-          "max_input_tokens": 32768,
-          "threshold_ms": 4000
-        },
-        {
-          "threshold_ms": 6000
-        }
-      ]
+      "kind": "linear_ms",
+      "intercept_ms": 4000,
+      "per_token_ms": 0.125
     },
     "tpot_rule": {
       "kind": "fixed_ms",
@@ -167,7 +156,7 @@
   },
   "search": {
     "low": 0.0,
-    "high": 0.25,
+    "high": 0.5,
     "tolerance": 0.001,
     "max_probes": 7,
     "sample_seed": 20260325