Pith. sign in

REVIEW 7 cited by

Arithmetic Without Algorithms: Language Models Solve Math With a Bag of Heuristics

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.21272 v2 pith:ZO25V2XL submitted 2024-10-28 cs.CL

classification cs.CL
keywords arithmeticheuristicneuronsalgorithmsheuristicsllmsmechanismmodel
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Do large language models (LLMs) solve reasoning tasks by learning robust generalizable algorithms, or do they memorize training data? To investigate this question, we use arithmetic reasoning as a representative task. Using causal analysis, we identify a subset of the model (a circuit) that explains most of the model's behavior for basic arithmetic logic and examine its functionality. By zooming in on the level of individual circuit neurons, we discover a sparse set of important neurons that implement simple heuristics. Each heuristic identifies a numerical input pattern and outputs corresponding answers. We hypothesize that the combination of these heuristic neurons is the mechanism used to produce correct arithmetic answers. To test this, we categorize each neuron into several heuristic types-such as neurons that activate when an operand falls within a certain range-and find that the unordered combination of these heuristic types is the mechanism that explains most of the model's accuracy on arithmetic prompts. Finally, we demonstrate that this mechanism appears as the main source of arithmetic accuracy early in training. Overall, our experimental results across several LLMs show that LLMs perform arithmetic using neither robust algorithms nor memorization; rather, they rely on a "bag of heuristics".

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Can Transformers Really Do It All? On the Compatibility of Inductive Biases Across Tasks

    cs.LG 2026-07 conditional novelty 7.0 of 10

    Learned replacement non-linearities show transformers are rarely optimal for algorithmic tasks, with benefits that are task-specific, while language/code gains are smaller and more transferable.

  2. Are Arithmetic Heuristic Neurons Form-Invariant? A Mechanistic Analysis of Symbols, Text, and Code in LLMs

    cs.CL 2026-07 conditional novelty 6.0 of 10

    Llama-3 arithmetic is computed by a shared neuron set across symbolic, text, and code formats, and cross-format failures are activation-state differences, not distinct circuits.

  3. All for One: LLMs Solve Mental Math at the Last Token With Information Transferred From Other Tokens

    cs.CL 2025-09 conditional novelty 6.0 of 10

    LLMs solve arithmetic in-context via an All-for-One pattern, with all input-specific computation occurring at the last token after a two-layer information transfer window.

  4. Modular Arithmetic: Language Models Solve Math Digit by Digit

    cs.CL 2025-08 conditional novelty 6.0 of 10

    LLMs perform 3-digit addition and subtraction via digit-position-specific MLP circuits that can be intervened upon to change individual output digits.

  5. Addition in Four Movements: Mapping Layer-wise Information Trajectories in LLMs

    cs.AI 2025-06 conditional novelty 5.0 of 10

    In LLaMA-3-8B-Instruct, addition information becomes linearly decodable in a fixed layer-wise order: formula structure, carries and sums, result digits, then output generation.

  6. Locate-then-Merge: Neuron-Level Parameter Fusion for Mitigating Catastrophic Forgetting in Multimodal LLMs

    cs.CL 2025-05 conditional novelty 5.0 of 10

    Neuron-Fusion selectively restores large-change neurons from a fine-tuned multimodal model and suppresses small changes, improving language retention with modest visual trade-offs.

  7. Reward Modeling for Reinforcement Learning-Based LLM Reasoning: Design, Challenges, and Evaluation

    cs.LG 2026-02 conditional novelty 4.0 of 10

    A taxonomy-driven survey arguing that reward design is the central mechanism shaping reliable LLM reasoning, with maps of reward paradigms, reward-hacking failure modes, and benchmark pitfalls.

Pith tools