SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2302.07863 · NeurIPS · 2023

Speculative Decoding with Big Little Decoder

Kurt Keutzer, Michael Mahoney, Jitendra Malik, Sehoon Kim, Amir Gholami, Karttikeya Mangalam, Suhong Moon

arXiv · PDF · Open in the Atlas

Code that ran

2 of 3 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 2 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2302.07863.json · badge: /api/ran/2302.07863.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
kssteven418/BigLittleDecoder canonical (classifier label)
classifier basis: classifier_2025
2 ran · 1 no recorded run
FunctionStatusWhere it lives
format_rouge_scores Ran
ran; our contract draft was wrong, not the code
kssteven418/BigLittleDecoder/examples/research_projects/bertabs/run_summarization.py
code served (permissive licence) · get_code("528b92cbc13c2b7b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

# Analyze the function to understand what structure 'scores' needs
# The function accesses:
# - scores['rouge-1']['f'], scores['rouge-1']['p'], scores['rouge-1']['r']
# - scores['rouge-2']['f'], scores['rouge-2']['p'], scores['rouge-2']['r']
# - scores['rouge-l']['f'], scores['rouge-l']['p'], scores['rouge-l']['r']
# So 'scores' is a dict with keys 'rouge-1', 'rouge-2', 'rouge-l'
# Each of those maps to a dict with keys 'f', 'p', 'r' (float values)

ARGS = {
    'scores': {
        'rouge-1': {
            'f': 0.752,
            'p': 0.845,
            'r': 0.681
        },
        'rouge-2': {
            'f': 0.623,
            'p': 0.714,
            'r': 0.556
        },
        'rouge-l': {
            'f': 0.698,
            'p': 0.789,
            'r': 0.621
        }
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
format_summary Ran
ran; our contract draft was wrong, not the code
kssteven418/BigLittleDecoder/examples/research_projects/bertabs/run_summarization.py
code served (permissive licence) · get_code("26e03f8ee35c78b3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function expects `translation` to be a tuple with at least 3 elements
# where the first element (raw_summary) is a string that will be processed
# with various replacements and stripping operations.

raw_summary = "[unused0]Hello[unused3] world[PAD][unused1] +test [unused2] example[unused2]"
translation = (raw_summary, None, None)

ARGS = {
    "translation": translation
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
build_data_iterator Not yet run
unverified (harvested, no recorded run)
kssteven418/BigLittleDecoder/examples/research_projects/bertabs/run_summarization.py
code served (permissive licence) · get_code("39dcd3b9f24258ff")
How it ran

no invocation recorded.

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 0 later) and 0 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
build_data_iterator 1 other paper
1 dated earlier
2106.13822 (2021-06)
XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Languages
format_rouge_scores 1 other paper
1 dated earlier
2106.13822 (2021-06)
XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Languages
format_summary 1 other paper
1 dated earlier
2106.13822 (2021-06)
XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Languages

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

The recent emergence of Large Language Models based on the Transformer architecture has enabled dramatic advancements in the field of Natural Language Processing. However, these models have long inference latency, which limits their deployment and makes them prohibitively expensive for various real-time applications. The inference latency is further exacerbated by autoregressive generative tasks, as models need to run iteratively to generate tokens sequentially without leveraging token-level parallelization. To address this, we propose Big Little Decoder (BiLD), a framework that can improve inference efficiency and latency for a wide range of text generation applications. The BiLD framework contains two models with different sizes that collaboratively generate text. The small model runs autoregressively to generate text with a low inference cost, and the large model is only invoked occasionally to refine the small model's inaccurate predictions in a non-autoregressive manner. To coordinate the small and large models, BiLD introduces two simple yet effective policies: (1) the fallback policy that determines when to hand control over to the large model; and (2) the rollback policy that determines when the large model needs to correct the small model's inaccurate predictions. To evaluate our framework across different tasks and models, we apply BiLD to various text generation scenarios encompassing machine translation on IWSLT 2017 De-En and WMT 2014 De-En, and summarization on XSUM and CNN/DailyMail. On an NVIDIA T4 GPU, our framework achieves a speedup of up to 2.12× speedup with minimal generation quality degradation. Furthermore, our framework is fully plug-and-play and can be applied without any modifications in the training process or model architecture. Our code is open-sourced 1 .

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2302.07863")
get_code_for_paper("2302.07863")
have("2302.07863")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2302.07863.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2302.07863.svg)](https://syntology.ai/paper/2302.07863)

Connect an agent — have() is free.