WEBDIENSTE/openswarm
1
0
Fork 0
mirror of https://github.com/openswarm-ai/openswarm.git synced 2026-08-31 20:29:56 +02:00
Code Issues Packages Projects Releases Wiki Activity

arena: v22 -- truncation made structurally impossible (action-first reply order)

Browse Source 
Per the design-out-failure-classes principle: the action line now LEADS the reply and
the PLAN trails as optional commentary, so a truncated reply can only lose commentary,
never the action -- the entire empty-action loss class becomes unrepresentable. The
bare-prompt retry stays as defense-in-depth behind it. v21 (patch-only variant) was
killed mid-sweep: the prompt edit contaminated its remaining tasks, and a mixed run is
not a measurement.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WsbS5x2rYsMDxP2kW3qqmQ
This commit is contained in:
ciregenz
2026-08-12 15:50:17 -07:00
co-authored by Claude Fable 5
parent ce418d4df6
commit 8cc2c99c32
1 changed files with 8 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
e2e/browser-v3/arena/llm_policy.py
+8 -2
View File
@@ -541,8 +541,8 @@ class OpenSwarmLlmPolicy(LlmPolicy):
# v5's addition, kept as its own constant so a supervisor restart never mutates v4 mid-sweep:
# browser-use's eval-memory loop won 35 multi-step tasks against v3's click-a-near-miss habit.
OSW_SYSTEM_V5 = OSW_SYSTEM + """
Start your reply with one short line: PLAN: <what is done, what remains, what to check next>.
Then the action on its own line. If the goal names a target you cannot see in the list or the page
Reply with your action call(s) FIRST, each on its own line. AFTER them you may add one line:
PLAN: <what is done, what remains, what to check next>. Never put anything before the first action. If the goal names a target you cannot see in the list or the page
text, EXPLORE first (switch tabs, scroll, open sections) -- never act on a near-miss. Before any
final submit/done click, re-check that every part of the goal is satisfied."""
@@ -704,6 +704,12 @@ def build(name: str, model: str = "", endpoint: str = "", **_: Any) -> Any:
scripted_drag=True, auto_complete=True, som=False,
native_pickers=True, verify_terminal=True, post_mouse_vision=True,
multi_cap=6, fill_verify=True, **v17)
if name == "osw-llm-v22": # STRUCTURAL truncation fix: action-first reply order (class impossible)
v22 = dict(v7, system=OSW_SYSTEM_V8 + OSW_SYSTEM_V9_WIDGETS + OSW_SYSTEM_V16, max_tokens=800)
return OpenSwarmLlmPolicy(name=name, multi=True, vision="progressive", fastpath=True,
scripted_drag=True, auto_complete=True, som=False,
native_pickers=True, verify_terminal=True, post_mouse_vision=True,
multi_cap=6, fill_verify=True, dispatch=True, offscreen=True, **v22)
if name == "osw-llm-v21": # v20 + bare-prompt retry (the 7 residual truncation deaths)
v21 = dict(v7, system=OSW_SYSTEM_V8 + OSW_SYSTEM_V9_WIDGETS + OSW_SYSTEM_V16, max_tokens=800)
return OpenSwarmLlmPolicy(name=name, multi=True, vision="progressive", fastpath=True,