SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.12865 · EMNLP · 2024

Backward Lens: Projecting Language Model Gradients into the Vocabulary Space

Lior Wolf, Yonatan Belinkov, Mor Geva, Shahar Katz

arXiv · PDF · Open in the Atlas

Code that ran

2 of 5 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; 3 no recorded run), 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/2402.12865.json · badge: /api/ran/2402.12865.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
shacharKZ/BackwardLens community
Papers with Code archive
2 ran · 3 no recorded run
FunctionStatusWhere it lives
auto_model_to_path_config Ran
ran; our contract draft was wrong, not the code
shacharKZ/BackwardLens/llm_utils.py
pointer only (licence: NONE) · get_code("9db9df78c98eebe2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, shacharKZ/BackwardLens 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.

extract_hs_include_prefix Ran
ran; our contract draft was wrong, not the code
shacharKZ/BackwardLens/llm_utils.py
pointer only (licence: NONE) · get_code("ea5da01bc0fb6cd3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, shacharKZ/BackwardLens 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.

GraphConfigs Not yet run
unverified (harvested, no recorded run)
shacharKZ/BackwardLens/llm_utils.py
pointer only (licence: NONE) · get_code("fb367523f3897fe7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, shacharKZ/BackwardLens 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. No invocation recorded.

auto_model_to_config Not yet run
unverified (harvested, no recorded run)
shacharKZ/BackwardLens/llm_utils.py
pointer only (licence: NONE) · get_code("b800598af3ba2e0c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, shacharKZ/BackwardLens 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. No invocation recorded.

wrap_model Not yet run
unverified (harvested, no recorded run)
shacharKZ/BackwardLens/llm_utils.py
pointer only (licence: NONE) · get_code("94c41d6c78d2a364")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, shacharKZ/BackwardLens 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. No invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 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.

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 how Transformer-based Language Models (LMs) learn and recall information is a key goal of the deep learning community. Recent interpretability methods project weights and hidden states obtained from the forward pass to the models' vocabularies, helping to uncover how information flows within LMs. In this work, we extend this methodology to LMs' backward pass and gradients. We first prove that a gradient matrix can be cast as a low-rank linear combination of its forward and backward passes' inputs. We then develop methods to project these gradients into vocabulary items and explore the mechanics of how new information is stored in the LMs' neurons. Our code is available at: https: //github.com/shacharKZ/BackwardLens .

For agents

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

get_harvested_code_for_paper("2402.12865")
get_code_for_paper("2402.12865")
have("2402.12865")

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

curl https://syntology.ai/api/ran/2402.12865.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.12865.svg)](https://syntology.ai/paper/2402.12865)

Connect an agent — have() is free.