GET/api/v1/evals/{runId}/hotspots

Feature-correlation hotspots for an eval run (T0-F round-6)

Surfaces the (feature, value) cells where failure rate is significantly higher than the run's baseline. Result reads like "failures correlate with model=gpt-4o-mini + prompt_length>2000 (cell 64% vs baseline 12%, n=137)." Statistical significance via two-proportion z-test; ranked by lift × log(n).

Authentication

Send Authorization: Bearer YOUR_API_KEY on every request. Generate API keys at /dashboard/api-keys.

Parameters

runId in pathrequired
string
minSupport in query
integer
minLift in query
number
topK in query
integer
maxComboSize in query
integer

Response

200 example

{
  "success": true
}

All status codes

200Hotspot report.
400(no description)
401(no description)
403(no description)
404(no description)

Code samples

cURL

curl -X GET \
  https://evalguard.ai/api/v1/evals/{runId}/hotspots \
  -H "Authorization: Bearer $EVALGUARD_API_KEY" \

TypeScript

import { EvalGuard } from "@evalguard/sdk";

const client = new EvalGuard({ apiKey: process.env.EVALGUARD_API_KEY });

const response = await client.request({
  method: "GET",
  path: "/api/v1/evals/{runId}/hotspots",
});
console.log(response);

Python

from evalguard import EvalGuard
import os

client = EvalGuard(api_key=os.environ["EVALGUARD_API_KEY"])

response = client.request(method="GET", path="/api/v1/evals/{runId}/hotspots")
print(response)

Go

package main

import (
	"context"
	"fmt"
	"os"

	"github.com/evalguard/evalguard-go"
)

func main() {
	client := evalguard.NewClient(os.Getenv("EVALGUARD_API_KEY"))
	resp, err := client.Request(context.Background(), "GET", "/api/v1/evals/{runId}/hotspots", nil)
	if err != nil { panic(err) }
	fmt.Println(resp)
}

Errors

400401403404

Other Evals endpoints