SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.01692 · ICLR · 2025

U-shaped and Inverted-U Scaling behind Emergent Abilities of Large Language Models

Tung-Yu Wu, Melody Lo

arXiv · PDF · Open in the Atlas

Code that ran

5 of 6 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; 1 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 3 failures of our instrument; 1 no recorded run); 1 of them was also run on the host, not in the jail, by the peer-agreement lane that set its level, measured 2026-11-22; 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/2410.01692.json · badge: /api/ran/2410.01692.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
tony10101105/ExpEmergence canonical (classifier label)
classifier basis: census
5 ran · 1 no recorded run
FunctionStatusWhere it lives
sigmoid Ran
ran, violated the contract we drafted
tony10101105/ExpEmergence/utils.py
code served (permissive licence) · get_code("b8e95809ca2c17c9")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned ndarray, ndim 1, dtype family float, shape (4,) as a later run recorded it. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-05T02:39:00Z). If your copy returns the same type, ndim and shape 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

# The sigmoid function takes a single parameter 'x' which should be a numeric value or array
# Looking at the function body, it performs: 1 / (1 + np.exp(-x))
# This works with scalars, lists, or numpy arrays

# Create a small numpy array to pass as input
x_input = np.array([0.0, 1.0, -1.0, 2.0])

ARGS = {
    "x": x_input
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (4,) as a later run recorded it
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
load_csv Ran
ran, honoured the contract we drafted
tony10101105/ExpEmergence/evaluation/abstract_narrative_understanding/abstract_narrative_understanding_question_grouping.py
code served (permissive licence) · get_code("dbdc88cccaa015e1")
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 DataFrame. 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 pandas as pd
import io
import tempfile
import os

# Create a minimal CSV file in memory and write it to a temporary file
csv_content = """col1,col2,col3
1,2,3
4,5,6
7,8,9"""

# Create a temporary file
temp_fd, temp_path = tempfile.mkstemp(suffix='.csv')
try:
    with os.fdopen(temp_fd, 'w') as f:
        f.write(csv_content)
    
    ARGS = {
        'file_path': temp_path
    }
finally:
    # Clean up will happen after execution
    pass
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 and dtype family int 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
model_filter Ran
ran; our contract draft was wrong, not the code
tony10101105/ExpEmergence/utils.py
code served (permissive licence) · get_code("1496edf0d82ed9cc")
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 DataFrame. 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 pandas as pd
import torch

torch.manual_seed(0)

# Create a sample dataframe with Model column and some other columns
df = pd.DataFrame({
    'Model': [
        'microsoft/phi-1_5',
        'microsoft/phi-2',
        'Qwen/Qwen1.5-0.5B',
        'Qwen/Qwen1.5-1.8B',
        'Qwen/Qwen1.5-4B',
        'stabilityai/stablelm-2-1_6b',
        'openai/gpt2',
        'facebook/opt-125m',
        'google/flan-t5-small'
    ],
    'accuracy': [0.75, 0.78, 0.72, 0.76, 0.80, 0.74, 0.81, 0.79, 0.77],
    'loss': [0.45, 0.42, 0.48, 0.44, 0.40, 0.46, 0.39, 0.41, 0.43]
})

ARGS = {
    'df': df
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
read_jsonl Ran
ran; our contract draft was wrong, not the code
tony10101105/ExpEmergence/evaluation/abstract_narrative_understanding/abstract_narrative_understanding_question_grouping.py
code served (permissive licence) · get_code("69f3426a987055c9")
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 json
import os

# Create a temporary JSONL file with test data
temp_file = tempfile.NamedTemporaryFile(mode='w', suffix='.jsonl', delete=False, encoding='utf-8')
temp_file_path = temp_file.name

# Write test JSONL data
test_data = [
    {"id": 1, "text": "hello world"},
    {"id": 2, "text": "foo bar"},
    {"id": 3, "text": "test data"}
]

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

ARGS = {
    "file_path": temp_file_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
whiten Ran
ran; our contract draft was wrong, not the code
tony10101105/ExpEmergence/utils.py
code served (permissive licence) · get_code("bc1e46e58d7a1d2d")
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 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

# Analyze the function: whiten(hex_color, k)
# - hex_color: a string in hex format (e.g., '#RRGGBB' or 'rgb(r,g,b)')
# - k: an integer to add to each color component
# The function expects hex_color to be parseable as either:
#   1. '#' followed by hex digits, or
#   2. 'rgb(r,g,b)' format
# Looking at the code: lstrip('#') removes '#', strip('rgb()') removes 'rgb()',
# then split(',') splits by comma
# This suggests the input should be in 'rgb(r,g,b)' format

# Testing the parsing logic:
# For 'rgb(100,150,200)':
# - lstrip('#') -> 'rgb(100,150,200)' (no '#' to remove)
# - strip('rgb()') -> '100,150,200'
# - split(',') -> ['100', '150', '200']

hex_color = 'rgb(100,150,200)'
k = 10

ARGS = {
    'hex_color': hex_color,
    'k': k
}
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 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
hex_to_rgba Not yet run
unverified (harvested, no recorded run)
tony10101105/ExpEmergence/plot_question_group_acc.py
code served (permissive licence) · get_code("4391f1331da022b9")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 5 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
sigmoid 51 other papers
44 dated earlier, 7 later · 2 only as an assembly brick
1505.05770 (2015-05)
Variational Inference with Normalizing Flows
1804.00779 (2018-04)
Neural Autoregressive Flows
1809.10765 (2018-09)
Auto-Encoding Knockoff Generator for FDR Controlled Variable Selection
and 48 more not named here

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

Large language models (LLMs) have been shown to exhibit emergent abilities in some downstream tasks, where model performance stagnates at first and then improves sharply and unpredictably with scale beyond a threshold. In this work, we investigate the phenomenon by grouping questions based on difficulty level and provide a possible explanation for emergent abilities. Specifically, we observe U-shaped scaling for hard questions and inverted-U scaling followed by steady improvement for easy questions. The two scaling patterns initially offset each other, causing stagnant overall performance. The performance starts to soar when the scaling pattern of easy questions reverts from inverse to standard scaling, leading to emergent abilities. Based on this finding, we propose a simple yet effective pipeline, called Slice-and-Sandwich, to predict the emergence threshold and model performance beyond the threshold. Our code is publicly available at https://github.com/tony10101105/ExpEmergence.

For agents

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

get_harvested_code_for_paper("2410.01692")
get_code_for_paper("2410.01692")
have("2410.01692")

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

curl https://syntology.ai/api/ran/2410.01692.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.01692.svg)](https://syntology.ai/paper/2410.01692)

Connect an agent — have() is free.