@arizeai/phoenix-evals 2.7.1 → 2.8.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (73) hide show
  1. package/CHANGELOG.md +6 -0
  2. package/README.md +40 -9
  3. package/dist/esm/decision/decideClassification.d.ts.map +1 -1
  4. package/dist/esm/decision/decideClassification.js +4 -4
  5. package/dist/esm/decision/decideClassification.js.map +1 -1
  6. package/dist/esm/llm/createCompletenessEvaluator.d.ts +1 -1
  7. package/dist/esm/llm/createCompletenessEvaluator.js +1 -1
  8. package/dist/esm/llm/createConcisenessEvaluator.d.ts +1 -1
  9. package/dist/esm/llm/createConcisenessEvaluator.js +1 -1
  10. package/dist/esm/llm/createCorrectnessEvaluator.d.ts +1 -1
  11. package/dist/esm/llm/createCorrectnessEvaluator.js +1 -1
  12. package/dist/esm/llm/createHallucinationEvaluator.d.ts +1 -1
  13. package/dist/esm/llm/createHallucinationEvaluator.js +1 -1
  14. package/dist/esm/llm/createPiiDetectionEvaluator.d.ts +1 -1
  15. package/dist/esm/llm/createPiiDetectionEvaluator.js +1 -1
  16. package/dist/esm/llm/createRefusalEvaluator.d.ts +1 -1
  17. package/dist/esm/llm/createRefusalEvaluator.js +1 -1
  18. package/dist/esm/llm/createRetrievalRelevanceEvaluator.d.ts +1 -1
  19. package/dist/esm/llm/createRetrievalRelevanceEvaluator.js +1 -1
  20. package/dist/esm/llm/createToolInvocationEvaluator.d.ts +1 -1
  21. package/dist/esm/llm/createToolInvocationEvaluator.js +1 -1
  22. package/dist/esm/llm/createToolResponseHandlingEvaluator.d.ts +1 -1
  23. package/dist/esm/llm/createToolResponseHandlingEvaluator.js +1 -1
  24. package/dist/esm/llm/createToolSelectionEvaluator.d.ts +1 -1
  25. package/dist/esm/llm/createToolSelectionEvaluator.js +1 -1
  26. package/dist/esm/llm/createToxicityEvaluator.d.ts +1 -1
  27. package/dist/esm/llm/createToxicityEvaluator.js +1 -1
  28. package/dist/esm/llm/createUserFrictionEvaluator.d.ts +1 -1
  29. package/dist/esm/llm/createUserFrictionEvaluator.js +1 -1
  30. package/dist/esm/tsconfig.esm.tsbuildinfo +1 -1
  31. package/dist/src/decision/decideClassification.d.ts.map +1 -1
  32. package/dist/src/decision/decideClassification.js +4 -4
  33. package/dist/src/decision/decideClassification.js.map +1 -1
  34. package/dist/src/llm/createCompletenessEvaluator.d.ts +1 -1
  35. package/dist/src/llm/createCompletenessEvaluator.js +1 -1
  36. package/dist/src/llm/createConcisenessEvaluator.d.ts +1 -1
  37. package/dist/src/llm/createConcisenessEvaluator.js +1 -1
  38. package/dist/src/llm/createCorrectnessEvaluator.d.ts +1 -1
  39. package/dist/src/llm/createCorrectnessEvaluator.js +1 -1
  40. package/dist/src/llm/createHallucinationEvaluator.d.ts +1 -1
  41. package/dist/src/llm/createHallucinationEvaluator.js +1 -1
  42. package/dist/src/llm/createPiiDetectionEvaluator.d.ts +1 -1
  43. package/dist/src/llm/createPiiDetectionEvaluator.js +1 -1
  44. package/dist/src/llm/createRefusalEvaluator.d.ts +1 -1
  45. package/dist/src/llm/createRefusalEvaluator.js +1 -1
  46. package/dist/src/llm/createRetrievalRelevanceEvaluator.d.ts +1 -1
  47. package/dist/src/llm/createRetrievalRelevanceEvaluator.js +1 -1
  48. package/dist/src/llm/createToolInvocationEvaluator.d.ts +1 -1
  49. package/dist/src/llm/createToolInvocationEvaluator.js +1 -1
  50. package/dist/src/llm/createToolResponseHandlingEvaluator.d.ts +1 -1
  51. package/dist/src/llm/createToolResponseHandlingEvaluator.js +1 -1
  52. package/dist/src/llm/createToolSelectionEvaluator.d.ts +1 -1
  53. package/dist/src/llm/createToolSelectionEvaluator.js +1 -1
  54. package/dist/src/llm/createToxicityEvaluator.d.ts +1 -1
  55. package/dist/src/llm/createToxicityEvaluator.js +1 -1
  56. package/dist/src/llm/createUserFrictionEvaluator.d.ts +1 -1
  57. package/dist/src/llm/createUserFrictionEvaluator.js +1 -1
  58. package/dist/tsconfig.tsbuildinfo +1 -1
  59. package/docs/classification.mdx +2 -2
  60. package/package.json +9 -9
  61. package/src/decision/decideClassification.ts +4 -5
  62. package/src/llm/createCompletenessEvaluator.ts +1 -1
  63. package/src/llm/createConcisenessEvaluator.ts +1 -1
  64. package/src/llm/createCorrectnessEvaluator.ts +1 -1
  65. package/src/llm/createHallucinationEvaluator.ts +1 -1
  66. package/src/llm/createPiiDetectionEvaluator.ts +1 -1
  67. package/src/llm/createRefusalEvaluator.ts +1 -1
  68. package/src/llm/createRetrievalRelevanceEvaluator.ts +1 -1
  69. package/src/llm/createToolInvocationEvaluator.ts +1 -1
  70. package/src/llm/createToolResponseHandlingEvaluator.ts +1 -1
  71. package/src/llm/createToolSelectionEvaluator.ts +1 -1
  72. package/src/llm/createToxicityEvaluator.ts +1 -1
  73. package/src/llm/createUserFrictionEvaluator.ts +1 -1
@@ -1 +1 @@
1
- {"version":3,"file":"decideClassification.d.ts","sourceRoot":"","sources":["../../../src/decision/decideClassification.ts"],"names":[],"mappings":"AAQA,OAAO,KAAK,EAAE,oBAAoB,EAAE,MAAM,gBAAgB,CAAC;AAC3D,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AACrD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AACnD,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,0BAA0B,CAAC;AAe9D;;;;;;;;GAQG;AACH,wBAAsB,oBAAoB,CAAC,IAAI,EAAE;IAC/C,KAAK,EAAE,aAAa,CAAC;IACrB,MAAM,EAAE,SAAS,MAAM,EAAE,CAAC;IAC1B,MAAM,EAAE,UAAU,CAAC;IACnB,SAAS,CAAC,EAAE,eAAe,CAAC;CAC7B,GAAG,OAAO,CAAC,oBAAoB,CAAC,CA2BhC"}
1
+ {"version":3,"file":"decideClassification.d.ts","sourceRoot":"","sources":["../../../src/decision/decideClassification.ts"],"names":[],"mappings":"AAQA,OAAO,KAAK,EAAE,oBAAoB,EAAE,MAAM,gBAAgB,CAAC;AAC3D,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AACrD,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AACnD,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,0BAA0B,CAAC;AAc9D;;;;;;;;GAQG;AACH,wBAAsB,oBAAoB,CAAC,IAAI,EAAE;IAC/C,KAAK,EAAE,aAAa,CAAC;IACrB,MAAM,EAAE,SAAS,MAAM,EAAE,CAAC;IAC1B,MAAM,EAAE,UAAU,CAAC;IACnB,SAAS,CAAC,EAAE,eAAe,CAAC;CAC7B,GAAG,OAAO,CAAC,oBAAoB,CAAC,CA2BhC"}
@@ -14,10 +14,10 @@ const decisionPrompt_1 = require("./decisionPrompt");
14
14
  */
15
15
  const QUESTION_ID = "label";
16
16
  /**
17
- * The question's instructions. The rendered prompt, rubric included, is sent
18
- * as the state.
17
+ * Instructions sent with the choice question. The OpenAI Decisions API rejects
18
+ * a choice question that has no instructions.
19
19
  */
20
- const INSTRUCTIONS = "Read the prompt in the state and answer it by selecting exactly one of the choices.";
20
+ const CHOICE_INSTRUCTIONS = "Select the single best label for the input.";
21
21
  /**
22
22
  * Classify with an AI SDK decision model by asking one choice question whose
23
23
  * options are the labels. Decision models don't generate text, so the
@@ -35,7 +35,7 @@ async function decideClassification(args) {
35
35
  questions: {
36
36
  [QUESTION_ID]: {
37
37
  type: "choice",
38
- instructions: INSTRUCTIONS,
38
+ instructions: CHOICE_INSTRUCTIONS,
39
39
  criteria: Object.fromEntries(labels.map((label) => [label, null])),
40
40
  },
41
41
  },
@@ -1 +1 @@
1
- {"version":3,"file":"decideClassification.js","sourceRoot":"","sources":["../../../src/decision/decideClassification.ts"],"names":[],"mappings":";;;AAAA;;;;GAIG;AACH,2BAAyC;AAEzC,4CAAmD;AAKnD,qDAAyD;AAEzD;;GAEG;AACH,MAAM,WAAW,GAAG,OAAO,CAAC;AAE5B;;;GAGG;AACH,MAAM,YAAY,GAChB,qFAAqF,CAAC;AAExF;;;;;;;;GAQG;AACI,KAAK,+BAA+B,IAK1C;IACC,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,IAAI,CAAC;IAClD,MAAM,MAAM,GAAG,MAAM,IAAA,wBAAmB,EAAC;QACvC,KAAK;QACL,KAAK,EAAE,IAAA,sCAAqB,EAAC,MAAM,CAAC;QACpC,SAAS,EAAE;YACT,CAAC,WAAW,CAAC,EAAE;gBACb,IAAI,EAAE,QAAQ;gBACd,YAAY,EAAE,YAAY;gBAC1B,QAAQ,EAAE,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;aACnE;SACF;QACD,SAAS,EAAE,IAAA,+BAAmB,EAAC;YAC7B,UAAU,EAAE,sBAAsB;YAClC,SAAS;SACV,CAAC;KACH,CAAC,CAAC;IACH,MAAM,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC;IAC3C,MAAM,UAAU,GAAG,aAAa,CAAC,MAAM,CAAC,gBAAgB,CAAC,CAAC;IAC1D,OAAO;QACL,KAAK,EAAE,MAAM,CAAC,MAAM;QACpB,QAAQ,gCACN,aAAa,EAAE,MAAM,CAAC,aAAa,IAChC,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,KACnD,OAAO,EAAE,MAAM,CAAC,QAAQ,CAAC,OAAO,GACjC;KACF,CAAC;AACJ,CAAC;AAED;;;;GAIG;AACH,SAAS,aAAa,CAAC,gBAAyB;;IAC9C,IAAI,OAAO,gBAAgB,KAAK,QAAQ,IAAI,gBAAgB,KAAK,IAAI,EAAE,CAAC;QACtE,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,KAAK,MAAM,QAAQ,IAAI,MAAM,CAAC,MAAM,CAAC,gBAAgB,CAAC,EAAE,CAAC;QACvD,MAAM,UAAU,GAAG,MAAC,QAAqD,aAArD,QAAQ,uBAAR,QAAQ,CACxB,UAAU,0CAAG,WAAW,CAAC,CAAC;QAC9B,IAAI,OAAO,UAAU,KAAK,QAAQ;YAAE,OAAO,UAAU,CAAC;IACxD,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}
1
+ {"version":3,"file":"decideClassification.js","sourceRoot":"","sources":["../../../src/decision/decideClassification.ts"],"names":[],"mappings":";;;AAAA;;;;GAIG;AACH,2BAAyC;AAEzC,4CAAmD;AAKnD,qDAAyD;AAEzD;;GAEG;AACH,MAAM,WAAW,GAAG,OAAO,CAAC;AAE5B;;;GAGG;AACH,MAAM,mBAAmB,GAAG,6CAA6C,CAAC;AAE1E;;;;;;;;GAQG;AACI,KAAK,+BAA+B,IAK1C;IACC,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,IAAI,CAAC;IAClD,MAAM,MAAM,GAAG,MAAM,IAAA,wBAAmB,EAAC;QACvC,KAAK;QACL,KAAK,EAAE,IAAA,sCAAqB,EAAC,MAAM,CAAC;QACpC,SAAS,EAAE;YACT,CAAC,WAAW,CAAC,EAAE;gBACb,IAAI,EAAE,QAAQ;gBACd,YAAY,EAAE,mBAAmB;gBACjC,QAAQ,EAAE,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;aACnE;SACF;QACD,SAAS,EAAE,IAAA,+BAAmB,EAAC;YAC7B,UAAU,EAAE,sBAAsB;YAClC,SAAS;SACV,CAAC;KACH,CAAC,CAAC;IACH,MAAM,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC;IAC3C,MAAM,UAAU,GAAG,aAAa,CAAC,MAAM,CAAC,gBAAgB,CAAC,CAAC;IAC1D,OAAO;QACL,KAAK,EAAE,MAAM,CAAC,MAAM;QACpB,QAAQ,gCACN,aAAa,EAAE,MAAM,CAAC,aAAa,IAChC,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,KACnD,OAAO,EAAE,MAAM,CAAC,QAAQ,CAAC,OAAO,GACjC;KACF,CAAC;AACJ,CAAC;AAED;;;;GAIG;AACH,SAAS,aAAa,CAAC,gBAAyB;;IAC9C,IAAI,OAAO,gBAAgB,KAAK,QAAQ,IAAI,gBAAgB,KAAK,IAAI,EAAE,CAAC;QACtE,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,KAAK,MAAM,QAAQ,IAAI,MAAM,CAAC,MAAM,CAAC,gBAAgB,CAAC,EAAE,CAAC;QACvD,MAAM,UAAU,GAAG,MAAC,QAAqD,aAArD,QAAQ,uBAAR,QAAQ,CACxB,UAAU,0CAAG,WAAW,CAAC,CAAC;QAC9B,IAAI,OAAO,UAAU,KAAK,QAAQ;YAAE,OAAO,UAAU,CAAC;IACxD,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}
@@ -34,7 +34,7 @@ export type CompletenessEvaluationRecord = {
34
34
  *
35
35
  * @example
36
36
  * ```ts
37
- * const evaluator = createCompletenessEvaluator({ model: openai("gpt-4o-mini") });
37
+ * const evaluator = createCompletenessEvaluator({ model: openai.decisionModel("gpt-6-luna") });
38
38
  * const result = await evaluator.evaluate({
39
39
  * conversation:
40
40
  * "User: Reset my password and update the billing address.\nAssistant: Your password has been reset.",
@@ -31,7 +31,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
31
31
  *
32
32
  * @example
33
33
  * ```ts
34
- * const evaluator = createCompletenessEvaluator({ model: openai("gpt-4o-mini") });
34
+ * const evaluator = createCompletenessEvaluator({ model: openai.decisionModel("gpt-6-luna") });
35
35
  * const result = await evaluator.evaluate({
36
36
  * conversation:
37
37
  * "User: Reset my password and update the billing address.\nAssistant: Your password has been reset.",
@@ -31,7 +31,7 @@ export type ConcisenessEvaluationRecord = {
31
31
  *
32
32
  * @example
33
33
  * ```ts
34
- * const evaluator = createConcisenessEvaluator({ model: openai("gpt-4o-mini") });
34
+ * const evaluator = createConcisenessEvaluator({ model: openai.decisionModel("gpt-6-luna") });
35
35
  * const result = await evaluator.evaluate({
36
36
  * input: "What is the capital of France?",
37
37
  * output: "Paris.",
@@ -32,7 +32,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
32
32
  *
33
33
  * @example
34
34
  * ```ts
35
- * const evaluator = createConcisenessEvaluator({ model: openai("gpt-4o-mini") });
35
+ * const evaluator = createConcisenessEvaluator({ model: openai.decisionModel("gpt-6-luna") });
36
36
  * const result = await evaluator.evaluate({
37
37
  * input: "What is the capital of France?",
38
38
  * output: "Paris.",
@@ -30,7 +30,7 @@ export type CorrectnessEvaluationRecord = {
30
30
  *
31
31
  * @example
32
32
  * ```ts
33
- * const evaluator = createCorrectnessEvaluator({ model: openai("gpt-4o-mini") });
33
+ * const evaluator = createCorrectnessEvaluator({ model: openai.decisionModel("gpt-6-luna") });
34
34
  * const result = await evaluator.evaluate({
35
35
  * input: "What is the capital of France?",
36
36
  * output: "Paris is the capital of France.",
@@ -31,7 +31,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
31
31
  *
32
32
  * @example
33
33
  * ```ts
34
- * const evaluator = createCorrectnessEvaluator({ model: openai("gpt-4o-mini") });
34
+ * const evaluator = createCorrectnessEvaluator({ model: openai.decisionModel("gpt-6-luna") });
35
35
  * const result = await evaluator.evaluate({
36
36
  * input: "What is the capital of France?",
37
37
  * output: "Paris is the capital of France.",
@@ -40,7 +40,7 @@ export interface HallucinationEvaluationRecord {
40
40
  *
41
41
  * @example
42
42
  * ```ts
43
- * const evaluator = createHallucinationEvaluator({ model: openai("gpt-4o-mini") });
43
+ * const evaluator = createHallucinationEvaluator({ model: openai.decisionModel("gpt-6-luna") });
44
44
  * const result = await evaluator.evaluate({
45
45
  * input:
46
46
  * "User: What's our refund window?\nTool (lookup_policy): Refunds: 30 days from delivery.\nAssistant: 30 days from delivery.\nUser: And for electronics?",
@@ -34,7 +34,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
34
34
  *
35
35
  * @example
36
36
  * ```ts
37
- * const evaluator = createHallucinationEvaluator({ model: openai("gpt-4o-mini") });
37
+ * const evaluator = createHallucinationEvaluator({ model: openai.decisionModel("gpt-6-luna") });
38
38
  * const result = await evaluator.evaluate({
39
39
  * input:
40
40
  * "User: What's our refund window?\nTool (lookup_policy): Refunds: 30 days from delivery.\nAssistant: 30 days from delivery.\nUser: And for electronics?",
@@ -34,7 +34,7 @@ export type PiiDetectionEvaluationRecord = {
34
34
  *
35
35
  * @example
36
36
  * ```ts
37
- * const evaluator = createPiiDetectionEvaluator({ model: openai("gpt-4o-mini") });
37
+ * const evaluator = createPiiDetectionEvaluator({ model: openai.decisionModel("gpt-6-luna") });
38
38
  * const result = await evaluator.evaluate({
39
39
  * conversation:
40
40
  * "User: Reset my account.\nAssistant: What email is on the account?\nUser: jane.doe@acme.com",
@@ -32,7 +32,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
32
32
  *
33
33
  * @example
34
34
  * ```ts
35
- * const evaluator = createPiiDetectionEvaluator({ model: openai("gpt-4o-mini") });
35
+ * const evaluator = createPiiDetectionEvaluator({ model: openai.decisionModel("gpt-6-luna") });
36
36
  * const result = await evaluator.evaluate({
37
37
  * conversation:
38
38
  * "User: Reset my account.\nAssistant: What email is on the account?\nUser: jane.doe@acme.com",
@@ -32,7 +32,7 @@ export type RefusalEvaluationRecord = {
32
32
  *
33
33
  * @example
34
34
  * ```ts
35
- * const evaluator = createRefusalEvaluator({ model: openai("gpt-4o-mini") });
35
+ * const evaluator = createRefusalEvaluator({ model: openai.decisionModel("gpt-6-luna") });
36
36
  * const result = await evaluator.evaluate({
37
37
  * input: "What is the capital of France?",
38
38
  * output: "I'm sorry, I can only help with technical questions.",
@@ -33,7 +33,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
33
33
  *
34
34
  * @example
35
35
  * ```ts
36
- * const evaluator = createRefusalEvaluator({ model: openai("gpt-4o-mini") });
36
+ * const evaluator = createRefusalEvaluator({ model: openai.decisionModel("gpt-6-luna") });
37
37
  * const result = await evaluator.evaluate({
38
38
  * input: "What is the capital of France?",
39
39
  * output: "I'm sorry, I can only help with technical questions.",
@@ -42,7 +42,7 @@ export interface RetrievalRelevanceEvaluationRecord {
42
42
  *
43
43
  * @example
44
44
  * ```ts
45
- * const evaluator = createRetrievalRelevanceEvaluator({ model: openai("gpt-4o-mini") });
45
+ * const evaluator = createRetrievalRelevanceEvaluator({ model: openai.decisionModel("gpt-6-luna") });
46
46
  * const result = await evaluator.evaluate({
47
47
  * input: "What is the capital of France?",
48
48
  * context: "Paris is the capital and largest city of France.",
@@ -35,7 +35,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
35
35
  *
36
36
  * @example
37
37
  * ```ts
38
- * const evaluator = createRetrievalRelevanceEvaluator({ model: openai("gpt-4o-mini") });
38
+ * const evaluator = createRetrievalRelevanceEvaluator({ model: openai.decisionModel("gpt-6-luna") });
39
39
  * const result = await evaluator.evaluate({
40
40
  * input: "What is the capital of France?",
41
41
  * context: "Paris is the capital and largest city of France.",
@@ -47,7 +47,7 @@ export type ToolInvocationEvaluationRecord = {
47
47
  *
48
48
  * @example
49
49
  * ```ts
50
- * const evaluator = createToolInvocationEvaluator({ model: openai("gpt-4o-mini") });
50
+ * const evaluator = createToolInvocationEvaluator({ model: openai.decisionModel("gpt-6-luna") });
51
51
  *
52
52
  * // Example with JSON schema format for available tools
53
53
  * const result = await evaluator.evaluate({
@@ -38,7 +38,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
38
38
  *
39
39
  * @example
40
40
  * ```ts
41
- * const evaluator = createToolInvocationEvaluator({ model: openai("gpt-4o-mini") });
41
+ * const evaluator = createToolInvocationEvaluator({ model: openai.decisionModel("gpt-6-luna") });
42
42
  *
43
43
  * // Example with JSON schema format for available tools
44
44
  * const result = await evaluator.evaluate({
@@ -47,7 +47,7 @@ export type ToolResponseHandlingEvaluationRecord = {
47
47
  *
48
48
  * @example
49
49
  * ```ts
50
- * const evaluator = createToolResponseHandlingEvaluator({ model: openai("gpt-4o-mini") });
50
+ * const evaluator = createToolResponseHandlingEvaluator({ model: openai.decisionModel("gpt-6-luna") });
51
51
  *
52
52
  * // Example: Correct extraction from tool result
53
53
  * const result = await evaluator.evaluate({
@@ -33,7 +33,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
33
33
  *
34
34
  * @example
35
35
  * ```ts
36
- * const evaluator = createToolResponseHandlingEvaluator({ model: openai("gpt-4o-mini") });
36
+ * const evaluator = createToolResponseHandlingEvaluator({ model: openai.decisionModel("gpt-6-luna") });
37
37
  *
38
38
  * // Example: Correct extraction from tool result
39
39
  * const result = await evaluator.evaluate({
@@ -48,7 +48,7 @@ export type ToolSelectionEvaluationRecord = {
48
48
  *
49
49
  * @example
50
50
  * ```ts
51
- * const evaluator = createToolSelectionEvaluator({ model: openai("gpt-4o-mini") });
51
+ * const evaluator = createToolSelectionEvaluator({ model: openai.decisionModel("gpt-6-luna") });
52
52
  *
53
53
  * const result = await evaluator.evaluate({
54
54
  * input: "User: What is the weather in San Francisco?",
@@ -39,7 +39,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
39
39
  *
40
40
  * @example
41
41
  * ```ts
42
- * const evaluator = createToolSelectionEvaluator({ model: openai("gpt-4o-mini") });
42
+ * const evaluator = createToolSelectionEvaluator({ model: openai.decisionModel("gpt-6-luna") });
43
43
  *
44
44
  * const result = await evaluator.evaluate({
45
45
  * input: "User: What is the weather in San Francisco?",
@@ -29,7 +29,7 @@ export type ToxicityEvaluationRecord = {
29
29
  *
30
30
  * @example
31
31
  * ```ts
32
- * const evaluator = createToxicityEvaluator({ model: openai("gpt-4o-mini") });
32
+ * const evaluator = createToxicityEvaluator({ model: openai.decisionModel("gpt-6-luna") });
33
33
  * const result = await evaluator.evaluate({
34
34
  * text: "You are a worthless idiot and everyone despises you.",
35
35
  * });
@@ -31,7 +31,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
31
31
  *
32
32
  * @example
33
33
  * ```ts
34
- * const evaluator = createToxicityEvaluator({ model: openai("gpt-4o-mini") });
34
+ * const evaluator = createToxicityEvaluator({ model: openai.decisionModel("gpt-6-luna") });
35
35
  * const result = await evaluator.evaluate({
36
36
  * text: "You are a worthless idiot and everyone despises you.",
37
37
  * });
@@ -36,7 +36,7 @@ export interface UserFrictionEvaluationRecord {
36
36
  *
37
37
  * @example
38
38
  * ```ts
39
- * const evaluator = createUserFrictionEvaluator({ model: openai("gpt-4o-mini") });
39
+ * const evaluator = createUserFrictionEvaluator({ model: openai.decisionModel("gpt-6-luna") });
40
40
  * const result = await evaluator.evaluate({
41
41
  * conversation: "User: Show recent orders.\nAssistant: Here are last month's orders.",
42
42
  * userMessage: "No, I asked for this week.",
@@ -32,7 +32,7 @@ const createClassificationEvaluator_1 = require("./createClassificationEvaluator
32
32
  *
33
33
  * @example
34
34
  * ```ts
35
- * const evaluator = createUserFrictionEvaluator({ model: openai("gpt-4o-mini") });
35
+ * const evaluator = createUserFrictionEvaluator({ model: openai.decisionModel("gpt-6-luna") });
36
36
  * const result = await evaluator.evaluate({
37
37
  * conversation: "User: Show recent orders.\nAssistant: Here are last month's orders.",
38
38
  * userMessage: "No, I asked for this week.",