State-writing models causally use edited scratchpad states in a controlled task at 80-91% accuracy on held-out examples, unlike final-answer-only and pretrained controls.
Arithmetic in the Wild: Llama uses Base-10 Addition to Reason About Cyclic Concepts
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
abstract
Does structure in representations imply structure in computation? We study how Llama-3.1-8B reasons over cyclic concepts (e.g., "what month is six months after August?"). Even though Llama-3.1-8B's representations for these concepts are circularly structured, we find that instead of directly computing modular addition in the period of the cyclic concept (e.g., 12 for months), the model re-uses a generic addition mechanism across tasks that operates independently of concept-specific geometry. First, it computes the sum of its two inputs using base-10 addition (six + August=14). Then, it maps this sum back to cyclic concept space (14->February). We show that Llama-3.1-8B uses task-agnostic Fourier features to compute these sums--in fact, these features have periods that respect standard base-10 addition, e.g., 2, 5, and 10, rather than the cyclic concept period (e.g., 12 for months). Furthermore, we identify a sparse set of 28 MLP neurons re-used across all tasks (approximately 0.2% of the MLP at layer 18) that can be partitioned into disjoint clusters, each computing the sum for a Fourier feature with a different period. Our work highlights how an interplay between causal abstraction and feature geometry can deepen our mechanistic understanding of LMs.
years
2026 4verdicts
UNVERDICTED 4representative citing papers
Single-neuron steering obeys a control-window law in which coherent behavior change occurs only when the trigger lies below a collapse ceiling computed from weights and one forward pass.
Transformer hidden states contain rank-indexed orientation signatures for true r-argument relations (r=3-6) that survive surface controls and can be patched to alter model outputs on relation tasks.
Larger models succeed on rare and complex tasks by reducing gradient interference from common tasks, allowing rare-task features to accumulate, as shown via synthetic task mixtures and OLMo pretraining from 4M to 4B parameters.
citing papers explorer
-
Do Models Read What They Write? Causal Registers in Scratchpad Reasoning
State-writing models causally use edited scratchpad states in a controlled task at 80-91% accuracy on held-out examples, unlike final-answer-only and pretrained controls.
-
Leverage Is Not Reach: A Control-Window Law for Single-Neuron Steering in Language Models
Single-neuron steering obeys a control-window law in which coherent behavior change occurs only when the trigger lies below a collapse ceiling computed from weights and one forward pass.
-
Relational Rank Geometry in Transformers: Detecting and Steering Hidden-State Relation Frames
Transformer hidden states contain rank-indexed orientation signatures for true r-argument relations (r=3-6) that survive surface controls and can be patched to alter model outputs on relation tasks.
-
Why Larger Models Learn More: Effects of Capacity, Interference, and Rare-Task Retention
Larger models succeed on rare and complex tasks by reducing gradient interference from common tasks, allowing rare-task features to accumulate, as shown via synthetic task mixtures and OLMo pretraining from 4M to 4B parameters.