SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2211.07443 · 2022

Calibrated Interpretation: Confidence Estimation in Semantic Parsing

arXiv · PDF · Open in the Atlas

Code that ran

7 of 9 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 6 have no contract verdict; 1 failure of our instrument; 2 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/2211.07443.json · badge: /api/ran/2211.07443.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
esteng/calibration_miso canonical (classifier label)
classifier basis not recorded on the link
6 ran · 2 no recorded run
esteng/calibration_metric canonical (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
get_and_tokenize Ran
ran on a synthesized input
esteng/calibration_miso/analysis/calibration_sql_lm.py
code served (permissive licence) · get_code("326bcd08f41f0633")
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, through the batch API) and returned list. 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 nltk
from nltk.lm.preprocessing import pad_sequence, padded_everygram_pipeline

class SimpleTokenizer:
    def tokenize(self, text):
        # Simple whitespace tokenizer
        return text.lower().split()

# Create sample data with 'plan' and 'utterance' keys
data = [
    {
        'plan': 'the quick brown fox',
        'utterance': 'a fast animal jumps'
    },
    {
        'plan': 'hello world example',
        'utterance': 'greetings to everyone'
    }
]

# Create a simple tokenizer
tokenizer = SimpleTokenizer()

# Define all arguments
ARGS = {
    'data': data,
    'key': 'plan',
    'tokenizer': tokenizer,
    'n': 5,
    'pad': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list output has no array shape of its own; none is 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
get_dataloader Ran
ran on a synthesized input
esteng/calibration_miso/analysis/calibration_sql_lstm_lm.py
code served (permissive licence) · get_code("7136734fee77a93e")
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, through the batch API) and returned DataLoader. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
import torch.nn as nn

torch.manual_seed(0)

# Build a small vocabulary dict
vocab = {
    'hello': 0,
    'world': 1,
    'foo': 2,
    'bar': 3,
    '<pad>': 4,
    '<eos>': 5
}

# Build small sentence examples (list of lists of tokens)
sents = [
    ['hello', 'world'],
    ['foo', 'bar', 'hello'],
    ['world']
]

ARGS = {
    'sents': sents,
    'vocab': vocab
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataLoader — a DataLoader output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataLoader 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
get_low_prob Ran
ran on a synthesized input
esteng/calibration_miso/analysis/calibration_ensemble.py
code served (permissive licence) · get_code("3b28e76018403525")
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, through the batch API) and returned tuple. 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 numpy as np

# Create an iterator that yields (idx, example) tuples
def create_iterator(is_miso=False):
    """Create a simple iterator for testing."""
    if is_miso:
        # MISO format: example[0] can be either a tuple with probs or a dict
        examples = [
            (0, ([0.8, 0.5, 0.3], "data")),  # tuple format with min_prob = 0.3
            (1, ({"expression_probs": [0.9, 0.7, 0.4]}, "data")),  # dict format with min_prob = 0.4
            (2, ([0.95, 0.92, 0.91], "data")),  # all high probs, min = 0.91
        ]
    else:
        # Non-MISO format: example has 'token_logprobs'
        # logprobs will be converted to probs via np.exp
        examples = [
            (0, {"token_logprobs": [np.array([-0.5, -1.0, -2.0])]}),  # probs ~ [0.606, 0.368, 0.135], min ~ 0.135
            (1, {"token_logprobs": [np.array([-0.1, -0.2, -0.15])]}),  # probs ~ [0.905, 0.819, 0.861], min ~ 0.819
            (2, {"token_logprobs": [np.array([-0.01, -0.02])]}),  # probs ~ [0.99, 0.98], min ~ 0.98
        ]
    
    for idx, example in examples:
        yield idx, example

ARGS = {
    "iterator": create_iterator(is_miso=False),
    "is_miso": False,
    "threshold": 0.6,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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
get_low_prob Ran
ran on a synthesized input
esteng/calibration_miso/analysis/calibration_make_data_splits.py
code served (permissive licence) · get_code("292f09d035dd264e")
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, through the batch API) and returned list, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 numpy as np

# Create iterator for non-MISO case
def create_iterator_non_miso():
    # Example with low probability (will trigger low_prob_idxs append)
    example1 = {
        'token_logprobs': [np.array([np.log(0.3), np.log(0.5), np.log(0.7)])]
    }
    # Example with high probability (will not trigger append)
    example2 = {
        'token_logprobs': [np.array([np.log(0.8), np.log(0.9), np.log(0.85)])]
    }
    return [(0, example1), (1, example2)]

# Create iterator for MISO case
def create_iterator_miso():
    # Example with low min_prob (tuple format)
    example1 = [(0.2, 0.2)]
    # Example with high min_prob (tuple format)
    example2 = [(0.9, 0.9)]
    return [(0, example1), (1, example2)]

# Non-MISO iterator
iterator_non_miso = create_iterator_non_miso()

# MISO iterator
iterator_miso = create_iterator_miso()

ARGS = {
    'iterator': iterator_non_miso,
    'is_miso': False,
    'threshold': 0.6
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
plot_df Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
esteng/calibration_metric/calibration_metric/vis/calibration_plot.py
pointer only (licence: NONE) · get_code("d2a31edb99a4fdfa")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, esteng/calibration_metric when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

read_data Ran
ran on a synthesized input
esteng/calibration_miso/analysis/calibration_sql_lm.py
code served (permissive licence) · get_code("162f3413a0413984")
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, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 json
import tempfile
import os

# Create a temporary file with JSON lines data
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.jsonl')
temp_path = temp_file.name

# Write some sample JSON lines to the file
sample_data = [
    {"id": 1, "text": "hello world"},
    {"id": 2, "text": "foo bar"},
    {"id": 3, "text": "test data"}
]

for item in sample_data:
    temp_file.write(json.dumps(item) + '\n')

temp_file.close()

ARGS = {
    'path': temp_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
read_test_data Ran
ran on a synthesized input
esteng/calibration_miso/analysis/calibration_sql_lm.py
code served (permissive licence) · get_code("fb231fe58ed02d5a")
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, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 tempfile
import os

# Create a temporary file with test data
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_file.write("line one\n")
temp_file.write("line two\n")
temp_file.write("line three\n")
temp_file.close()

ARGS = {
    'path': temp_file.name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
get_low_prob_idxs Not yet run
unverified (harvested, no recorded run)
esteng/calibration_miso/analysis/calibration_make_data_splits.py
code served (permissive licence) · get_code("3c84fab5b11ff0ff")
How it ran

no invocation recorded.

get_ppl Not yet run
unverified (harvested, no recorded run)
esteng/calibration_miso/analysis/calibration_sql_lstm_lm.py
code served (permissive licence) · get_code("679c1b4f3abe2d88")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 9 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.

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

Sequence generation models are increasingly being used to translate natural language into programs, i.e. to perform executable semantic parsing. The fact that semantic parsing aims to predict programs that can lead to executed actions in the real world motivates developing safe systems. This in turn makes measuring calibration -- a central component to safety -- particularly important. We investigate the calibration of popular generation models across four popular semantic parsing datasets, finding that it varies across models and datasets. We then analyze factors associated with calibration error and release new confidence-based challenge splits of two parsing datasets. To facilitate the inclusion of calibration in semantic parsing evaluations, we release a library for computing calibration metrics.

For agents

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

get_harvested_code_for_paper("2211.07443")
get_code_for_paper("2211.07443")
have("2211.07443")

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

curl https://syntology.ai/api/ran/2211.07443.json

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

[![Syntology run record](https://syntology.ai/api/ran/2211.07443.svg)](https://syntology.ai/paper/2211.07443)

Connect an agent — have() is free.