Exact input contract for an independent eager Hugging Face next-token comparison. It binds the source revision, model bytes, rendered prompt, continuation prefix, and disputed Kiln and vLLM candidates; the oracle result is recorded separately.
Schema identity
| Property | Value |
|---|---|
| Title | Kiln Hugging Face Next-Token Request v1 |
$id |
https://ericflo.github.io/kiln/schemas/hf-next-token-request-v1.schema.json |
| Dialect | https://json-schema.org/draft/2020-12/schema |
| Root type | object |
| Root object | closed |
Root fields
| Field | Required | Type | Constraints and default | Description |
|---|---|---|---|---|
schema |
yes | any |
const "kiln.hf-next-token-request.v1" | Schema identifier for this request format. |
id |
yes | string |
pattern ^[a-z0-9][a-z0-9._-]{2,127}$ | Stable, path-safe identifier for this attribution request. |
source |
yes | object |
closed object | Source revision and qualification receipts that produced the two disputed candidate tokens. |
model_identity |
yes | modelIdentity |
- | Content identity of the exact model, tokenizer, chat template, configuration, and weight shards used for attribution. |
prompt |
yes | object |
closed object | Single-user prompt recipe and the exact rendered token sequence before the continuation prefix is appended. |
continuation_prefix |
yes | array<token> |
minimum items 1; maximum items 64 | Ordered token sequence forced after the rendered prompt to reproduce the first divergent next-token state. |
input_token_ids |
yes | tokenIds |
- | Complete oracle input: the rendered prompt token IDs followed by the continuation-prefix token IDs. |
input_token_ids_sha256 |
yes | sha256 |
- | SHA-256 identity of the canonical complete input-token sequence. |
candidates |
yes | array |
minimum items 2; maximum items 2 | Exactly two disputed next-token candidates, ordered as Kiln first and vLLM second. |
Definitions
sha256Lowercase SHA-256 digest with the `sha256:` algorithm prefix.0 fields
Type: string. Constraints: pattern ^sha256:[0-9a-f]{64}$.
tokenIdTokenizer vocabulary ID for the bound Qwen3.5-4B model identity.0 fields
Type: integer. Constraints: minimum 0; maximum 248319.
tokenIdsBounded, nonempty sequence of tokenizer vocabulary IDs.0 fields
Type: array<tokenId>. Constraints: minimum items 1; maximum items 4096.
tokenOne continuation token represented by both vocabulary ID and decoded text.2 fields
Type: object. Constraints: closed object.
| Field | Required | Type | Constraints and default | Description |
|---|---|---|---|---|
token_id |
yes | tokenId |
- | Vocabulary ID of this continuation token. |
text |
yes | string |
- | Tokenizer-decoded text for this token; it may be empty or contain nonprinting content. |
candidateOne engine's disputed next-token prediction at the bound prompt and continuation state.3 fields
Type: object. Constraints: closed object.
| Field | Required | Type | Constraints and default | Description |
|---|---|---|---|---|
engine |
yes | any |
enum "kiln", "vllm" | Candidate-producing inference engine. |
token_id |
yes | tokenId |
- | Vocabulary ID predicted by the named engine. |
text |
yes | string |
- | Tokenizer-decoded text for the predicted token. |
receiptReferenceRepository-relative qualification receipt with hashes of both its file bytes and canonical content.3 fields
Type: object. Constraints: closed object.
| Field | Required | Type | Constraints and default | Description |
|---|---|---|---|---|
path |
yes | string |
minimum length 1 | Repository-relative path to the retained receipt. |
file_sha256 |
yes | sha256 |
- | SHA-256 identity of the receipt file's exact bytes. |
content_sha256 |
yes | sha256 |
- | SHA-256 identity of the receipt's canonical content, excluding incidental serialization differences. |
weightFileOne model weight shard and its exact size and content identity.3 fields
Type: object. Constraints: closed object.
| Field | Required | Type | Constraints and default | Description |
|---|---|---|---|---|
path |
yes | string |
minimum length 1 | Path or stable shard name within the bound model source. |
bytes |
yes | integer |
minimum 1 | Exact weight-shard size in bytes. |
sha256 |
yes | sha256 |
- | SHA-256 identity of the shard's exact bytes. |
modelIdentityComplete content identity for the model inputs used by every candidate and oracle process.6 fields
Type: object. Constraints: closed object.
| Field | Required | Type | Constraints and default | Description |
|---|---|---|---|---|
id |
yes | string |
minimum length 1 | Human-readable model identifier recorded alongside the content hashes. |
content_sha256 |
yes | sha256 |
- | Aggregate content identity for the complete model source. |
config_hash |
yes | sha256 |
- | SHA-256 identity of the effective model configuration. |
tokenizer_hash |
yes | sha256 |
- | SHA-256 identity of the tokenizer inputs. |
chat_template_hash |
yes | sha256 |
- | SHA-256 identity of the chat template used to render the prompt. |
weight_files |
yes | array<weightFile> |
minimum items 1 | Complete ordered inventory of model weight shards. |