Token-level eval harness for tool-calling agents: what we wired up

May 26, 2026 · 4:03 PM UTC ·4 min read · 0 reactions · 0 comments · 35 views

TL;DR · WeSearch summary

Nexus Labs has developed a token-level evaluation harness for tool-calling agents to improve performance assessment. This new system replaces a simple pass/fail metric with four distinct signals that provide deeper insights into agent behavior. The implementation aims to enhance the accuracy of tool selection and argument handling, addressing issues that previously went unnoticed in evaluations.

Key facts

▪The evaluation harness now measures tool selection accuracy, argument F1, recovery rate, and trajectory length delta separately.
▪Previously, a single pass rate of 72% did not reveal issues with tool selection and argument handling, prompting the redesign.
▪The new system allows for running evaluations against multiple models without rewriting the harness, using a single OpenAI-compatible endpoint.

About this source

DEV.to (Top) files mainly under programming. We currently carry 4,924 of its stories.

All DEV.to (Top) coverage →

Original article

DEV.to (Top)

Read full at DEV.to (Top) →

Story provenance

Source · retrieval · rights · ranking — open for full record

inspect →

Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.

Record

Original publisher	DEV.to (Top)
Canonical URL	https://dev.to/marcuswwchen/token-level-eval-harness-for-tool-calling-agents-what-we-wired-up-1m1b
Publication time	Tue, 26 May 2026 16:03:35 +0000
Retrieval time	2026-05-26T16:07:49.955Z
Last seen	2026-05-26T16:07:49.955Z
Headline source	Publisher (no WeSearch rewrite)
Excerpt source	publisher body
Excerpt method	First ~120 words (~800 chars) of extracted publisher body, fair-use limited.
Summary	WeSearch · cerebras-chat (WeSearch summarizer)
Summary source text	contentText
Citation coverage	Summary is a WeSearch-generated derivative; primary citation is the original publisher URL.
Cluster	FNcf-P3UVuDN
Cluster logic	Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison.
Ranking reason	Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking.
Publisher visit	Yes — open original
Substitutes article?	No — link-out required for full text

Rights status (four layers)

Publisher-declared

No publisher-confirmed rights record for this source yet.

Machine-readable

No source-specific machine-readable restriction detected beyond the public feed.

WeSearch interpretation

WeSearch declared handling (basis: Derived from the published RSS/Atom feed). This is WeSearch policy, not a legal grant on the publisher's behalf.

Unknown

Retrieval and training permissions are not asserted unless the publisher confirms them.

WeSearch handling by dimension

Indexing	May the item be indexed (stored, ranked, made findable)?	Allowed
Snippet	May a short excerpt of the publisher's text be shown?	Allowed
AI summary	May WeSearch generate its own short summary of the article?	Limited
Retrieval / RAG	May the content be exposed for third-party retrieval-augmented generation?	Not asserted
Model training	May the content be used to train AI models?	Not asserted
Commercial reuse	May the content be reused commercially?	Not permitted

Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.

Open original publisher → Source rights JSON → Machine-readable provenance → Rights framework → WESEARCH-STORY-PROVENANCE-V1

Opening excerpt (first ~120 words) tap to expand

try { if(localStorage) { let currentUser = localStorage.getItem('current_user'); if (currentUser) { currentUser = JSON.parse(currentUser); if (currentUser.id === 3859428) { document.getElementById('article-show-container').classList.add('current-user-is-article-author'); } } } } catch (e) { console.error(e); } Marcus Chen Posted on May 26 Token-level eval harness for tool-calling agents: what we wired up #mlops #llm #machinelearning #devops TL;DR: We replaced our "did the agent finish the task" pass/fail eval with a token-level harness that scores tool selection, argument shape, and recovery behavior separately. Pass rate went from a single 73% number to four signals that actually tell us what broke.

…

Excerpt limited to ~120 words for fair-use compliance. The full article is at DEV.to (Top).

Anonymous · no account needed

Discussion

0 comments