Skip to content

Commit 359c6ae

Browse files
gsvigruhaclaudeestherk15
authored
Document 250 KB per-span field size limit for LLM-as-a-judge evaluations (#35773)
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> Co-authored-by: Esther Kim <esther.kim@datadoghq.com>
1 parent 8525712 commit 359c6ae

1 file changed

Lines changed: 2 additions & 0 deletions

File tree

  • content/en/llm_observability/evaluations/custom_llm_as_a_judge_evaluations

content/en/llm_observability/evaluations/custom_llm_as_a_judge_evaluations/_index.md

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -277,6 +277,8 @@ Assessment Criteria is not currently available for JSON evaluations.
277277

278278
### Define the evaluation scope: Filtering and sampling
279279

280+
<div class="alert alert-info">Span fields used in evaluations are limited to 250 KB each. Fields exceeding this size are truncated before being sent to the LLM judge.</div>
281+
280282
Under **Evaluation Scope**, define where and how your evaluation runs. This helps control coverage (which spans are included) and cost (how many spans are sampled).
281283
- **Application**: Select the application you want to evaluate.
282284
- **Evaluate On**: Choose one of the following:

0 commit comments

Comments
 (0)