SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2209.02535 · 2022

Analyzing Transformers in Embedding Space

arXiv · PDF · Open in the Atlas

Code that ran

3 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 3 have no contract verdict; 0 failures of our instrument; 0 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/2209.02535.json · badge: /api/ran/2209.02535.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
guyd1995/embedding-space canonical (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
FunctionStatusWhere it lives
convert_to_tokens Ran
ran on a synthesized input
guyd1995/embedding-space/utils.py
code served (permissive licence) · get_code("5c671bb6e1706d40")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Stand-in tokenizer class that implements convert_ids_to_tokens
class MockTokenizer:
    def __init__(self):
        self.vocab = {
            0: 'hello',
            1: 'Ġworld',
            2: 'Ġtest',
            3: '#ing',
            4: 'Ġfoo',
        }
    
    def convert_ids_to_tokens(self, indices):
        """Convert token IDs to token strings"""
        return [self.vocab.get(idx, f'token_{idx}') for idx in indices]

# Create test data
indices = [0, 1, 2, 3, 4]
tokenizer = MockTokenizer()

ARGS = {
    'indices': indices,
    'tokenizer': tokenizer,
    'strip': True,
    'width': 15,
}
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 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
keep_k Ran
ran on a synthesized input
guyd1995/embedding-space/utils.py
code served (permissive licence) · get_code("0197737303f2577c")
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 Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. 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 torch

torch.manual_seed(0)

# Create a small tensor with distinct dimensions
# Shape: (2, 4) - batch size 2, feature dimension 4
x = torch.randn(2, 4)

# k parameter: number of top elements to keep
k = 2

# absolute parameter: whether to use absolute values for topk
absolute = True

# dim parameter: dimension along which to apply topk
dim = -1

ARGS = {
    "x": x,
    "k": k,
    "absolute": absolute,
    "dim": dim,
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 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
top_tokens Ran
ran on a synthesized input
guyd1995/embedding-space/utils.py
code served (permissive licence) · get_code("62e9095cb473e6ff")
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 torch
import numpy as np

torch.manual_seed(0)

# Create a minimal tokenizer stand-in with required attributes/methods
class MinimalTokenizer:
    def __init__(self):
        # vocab: maps token string -> token id
        self.vocab = {
            'hello': 0,
            'world': 1,
            'Ġtest': 2,
            'Ġfoo': 3,
            '#bar': 4,
            'Ġ[unused]': 5,
            'café': 6,
            'a': 7,
            'b': 8,
            'c': 9,
        }
    
    def convert_ids_to_tokens(self, indices):
        # Convert token ids back to token strings
        id_to_token = {v: k for k, v in self.vocab.items()}
        return [id_to_token.get(int(idx), '[UNK]') for idx in indices]

tokenizer = MinimalTokenizer()

# Create a small vector with scores for each token
# Shape should match vocab size
v = torch.randn(len(tokenizer.vocab))

ARGS = {
    'v': v,
    'tokenizer': tokenizer,
    'k': 5,
    'only_english': False,
    'only_ascii': False,
    'exclude_brackets': False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — 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

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, 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 3 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

Understanding Transformer-based models has attracted significant attention, as they lie at the heart of recent technological advances across machine learning. While most interpretability methods rely on running models over inputs, recent work has shown that a zero-pass approach, where parameters are interpreted directly without a forward/backward pass is feasible for some Transformer parameters, and for two-layer attention networks. In this work, we present a theoretical analysis where all parameters of a trained Transformer are interpreted by projecting them into the embedding space, that is, the space of vocabulary items they operate on. We derive a simple theoretical framework to support our arguments and provide ample evidence for its validity. First, an empirical analysis showing that parameters of both pretrained and fine-tuned models can be interpreted in embedding space. Second, we present two applications of our framework: (a) aligning the parameters of different models that share a vocabulary, and (b) constructing a classifier without training by ``translating'' the parameters of a fine-tuned classifier to parameters of a different model that was only pretrained. Overall, our findings open the door to interpretation methods that, at least in part, abstract away from model specifics and operate in the embedding space only.

For agents

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

get_harvested_code_for_paper("2209.02535")
get_code_for_paper("2209.02535")
have("2209.02535")

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

curl https://syntology.ai/api/ran/2209.02535.json

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

[![Syntology run record](https://syntology.ai/api/ran/2209.02535.svg)](https://syntology.ai/paper/2209.02535)

Connect an agent — have() is free.