arXiv subjects

Subhabrata Dutta

Publications and source records attributed to Subhabrata Dutta.

At least 19 recordsLinked to original sources

arXiv2026

MURANO: Design, Run, and Reproduce Mechanistic Interpretability Experiments as Composable Pipelines

This paper presents Murano, an open source framework for designing, running, and reproducing mechanistic interpretability studies of large language models, intended for researchers across disciplines. These studies often combine loading, recording, attribution, intervention, and evaluation, while existing libraries tend to focus on different parts of this workflow. As a result, researchers using several libraries may need to adapt outputs from one for use by another. To bridge this gap, Murano represents operations from these five areas as composable steps. Steps exchange named result artifacts and declare the inputs they require and the outputs they produce. A pipeline executes its steps in the order supplied, and Murano uses canonical addresses when component identities pass between operations. Murano builds on existing interpretability and machine learning libraries. We demonstrate Murano through two reproductions of established interpretability studies and one illustrative sparse autoencoder case study.

cs.CL↗

arXiv2026

Many Circuits, One Mechanism: Input Variation and Evaluation Granularity in Circuit Discovery

Circuit discovery methods identify subgraphs that explain model behaviors, and structural differences between discovered circuits are commonly interpreted as evidence of distinct mechanisms. We test this assumption by varying input-token frequency while holding the task fixed. The discovered circuits appear specialized by frequency when compared structurally, but functional and representational analyses show no reliable evidence of corresponding differences. We term this mismatch phantom specialization. Using the Literal Sequence Copying task across four frequency bands plus a frequency-weighted control, we extract 75 circuits from five Pythia models (70M-1.4B). We find that structurally distinct circuits implement the same computation: band-specific edges transfer broadly across bands, a core shared across most bands recovers at least 99% of circuit performance in models above 70M, and causal interchange interventions confirm that internal representations are interchangeable across frequency bands. A smaller subject-verb agreement replication shows the same pattern: circuits differ structurally but transfer broadly across bands, and the core shared by most bands recovers nearly all circuit accuracy. Repeated extractions within the same band further suggest that discovery algorithms sample from an equivalence class of valid subgraphs rather than recovering a unique mechanism. Standard evaluation practice obscures this pattern: source-level evaluation inflates apparent faithfulness, while edge-level evaluation reveals the many-to-one mapping from structure to function. We do not test specialization at the level of token positions or features inside a component. Our results show that structural differences between circuits are not sufficient evidence for distinct mechanisms, and that exposing this requires edge-level evaluation and cross-condition transfer tests.

cs.CL↗

arXiv2026

Sparse Autoencoders Encode Both Concepts and Functions: The Downstream Geometry of Feature Effects

The wide-scale use of sparse autoencoders (SAEs) as interpretability tools is limited by inconsistent links between SAE features and model behavior. Features with clear activation descriptions may have weak or unexpected causal effects; steering can vary across prompts or oppose the intended direction; and activation-based feature selection can miss features that produce the desired output change. Prior work has studied feature geometry inside the model, where features are computed. We instead study the geometry of changes in model logits caused by feature interventions. We introduce Feature-Effect Geometry Analysis (FEGA), an unsupervised framework that removes the same active SAE feature across contexts and analyzes the resulting cloud of logit changes. Across SAE variants, consistent one-dimensional effects are rare: few features behave like reusable directions. To interpret this variation, we distinguish value-like features, tied to static information such as factual attributes, from pointer-like features, associated with context-dependent operations. Value-like features more often exhibit structured, low-dimensional effects, although these effects typically span several directions. Pointer-like features, by contrast, predominantly exhibit diffuse effects. Our results show that a feature can be interpretable and causally relevant without providing a stable direction for steering.

cs.LG↗

arXiv2026

Hierarchical Latent Structures in Data Generation Process Unify Mechanistic Phenomena across Scale

Contemporary studies in mechanistic interpretability have uncovered many puzzling phenomena in the neural information processing of Transformer-based language models, such as induction heads, function vectors, and the Hydra effect. Some of these individual phenomena have been independently tied to different data distributional properties, while some have been loosely associated with model architecture and how Transformers process information. However, a unified understanding of the relationship between data, model architecture, and optimization remains lacking, failing to answer the fundamental question: why do these three phenomena appear universally across different model families and scales, despite their seeming disconnect? In this work, we answer this question by unifying these three phenomena as consequences of hierarchical latent structures in the data generation process, coupled with decorrelated gradients across additive model components and directional concavity in the representation geometry. We validate our theoretical results in a toy model regime and in a large-scale synthetic data regime, comparing them with language models trained on natural language data.

cs.CL↗

arXiv2026

The FIL Hypothesis: Inductive Biases Help with Kernel Engineering

The Bitter Lesson, which posits that general-purpose methods that scale with computation and data ultimately outperform those with built-in human knowledge, has become a dominant paradigm in the era of Large Language Models. We revisit this principle by observing a new and critical scaling dimension: the duration of the Feedback Information Loop (FIL), the time required for a system to receive a verification signal after generating a prediction. Most historic successes in Artificial Intelligence (AI) have benefited from near instantaneous feedback (e.g., games or classification tasks), but we argue that future AI applications in science and the physical world will inherently involve FILs ranging from hours to weeks. This trend poses a fundamental scaling limit, as obtaining enough verification steps required by purely data-driven methods becomes practically impossible. Additionally, we propose a method that is orthogonal to purely data-driven approaches, based on human-inspired expert knowledge. The method relies on inductive biases and constraining the solution space. We provide an initial validation of the hypothesis and the method, by studying the real-world GPU programming task, a domain with non-trivial FIL, and demonstrate that incorporating inductive biases yields superior performance over data-driven approaches. The code is released under: https://github.com/ai-nikolai/robust_kernelbench

cs.AI↗

arXiv2026

Expert Preference-based Evaluation of Automated Related Work Generation

Expert domain writing, such as scientific writing, typically demands extensive domain knowledge. Although large language models (LLMs) show promising potential in this task, evaluating the quality of automatically generated scientific writing is a crucial open issue, as it requires knowledge of domain-specific criteria and the ability to discern expert preferences. Conventional automatic evaluation metrics and LLM-as-a-judge systems, primarily designed for mainstream NLP tasks, are insufficient to grasp expert preferences and domain-specific quality standards. To address this gap and support realistic human-AI collaborative writing, we focus on related work generation, one of the most challenging scientific tasks, as an exemplar. We propose GREP, a multi-turn evaluation framework that integrates classical related work evaluation criteria with expert-specific preferences. GREP decomposes the evaluation into smaller fine-grained dimensions. This localized evaluation is further augmented with contrastive examples to provide detailed contextual guidance for the evaluation dimensions. Empirical investigation reveals that GREP is able to assess the quality of related work sections in a much more robust manner compared to standard LLM judges, reflects natural scenarios of scientific writing, and bears a strong correlation with the assessment of human experts. We also observe that generations from state-of-the-art LLMs struggle to satisfy validation constraints of a suitable related work section.

cs.CL↗

arXiv2026

Patches of Nonlinearity: Instruction Vectors in Large Language Models

Despite the recent success of instruction-tuned language models and their ubiquitous usage, very little is known of how models process instructions internally. In this work, we address this gap from a mechanistic point of view by investigating how instruction-specific representations are constructed and utilized in different stages of post-training: Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO). Via causal mediation, we identify that instruction representation is fairly localized in models. These representations, which we call Instruction Vectors (IVs), demonstrate a curious juxtaposition of linear separability along with non-linear causal interaction, broadly questioning the scope of the linear representation hypothesis commonplace in mechanistic interpretability. To disentangle the non-linear causal interaction, we propose a novel method to localize information processing in language models that is free from the implicit linear assumptions of patching-based techniques. We find that, conditioned on the task representations formed in the early layers, different information pathways are selected in the later layers to solve that task, i.e., IVs act as circuit selectors.

cs.CL↗

arXiv2026

Subjective Code Preferences in Experts and Large Language Models

Large Language Models (LLMs) have become increasingly popular for coding tasks, with subjective coding preferences being an essential element to adapt to programmers' personal needs. Existing work overlooks such characteristics and mainly focuses on code correctness. In this study, we propose a typification of four subjective coding preference axes - complexity, commenting, modularity, and readability - motivated by common engineering habits and validated by 25 software engineers. We collect a dataset of ~3,000 paired Python code snippets reflecting these axes, annotated by 73 experts who rate their preferences on a Likert scale. Using our dataset, we study how LLMs handle subjective coding preferences. We present 13 LLMs with pairs of solutions to the same programming task, first as textual descriptions and then as concrete code snippets. We find that models often prefer one option in natural language but the opposite when evaluating code. More consistent models (i.e., those that are coherent in their choices between deeds and words) frequently reveal positional bias: swapping the order of options changes the preferred alternative. We then use the five most consistent models to re-annotate the dataset. Compared to humans, models show polarized Likert distributions and notable divergence in ratings. A case study on GPT-5 reveals reliance on external assumptions and brittle reasoning.

cs.HC↗

arXiv2026

Reward Modeling for Scientific Writing Evaluation

Scientific writing is an expert-domain task that demands deep domain knowledge, task-specific requirements and reasoning capabilities that leverage the domain knowledge to satisfy the task specifications. While scientific text generation has been widely studied, its evaluation remains a challenging and open problem. It is critical to develop models that can be reliably deployed for evaluating diverse open-ended scientific writing tasks while adhering to their distinct requirements. However, existing LLM-based judges and reward models are primarily optimized for general-purpose benchmarks with fixed scoring rubrics and evaluation criteria. Consequently, they often fail to reason over sparse knowledge of scientific domains when interpreting task-dependent and multi-faceted criteria. Moreover, fine-tuning for each individual task is costly and impractical for low-resource settings. To bridge these gaps, we propose cost-efficient, open-source reward models tailored for scientific writing evaluation. We introduce a two-stage training framework that initially optimizes scientific evaluation preferences and then refines reasoning capabilities. Our multi-aspect evaluation design and joint training across diverse tasks enable fine-grained assessment and robustness to dynamic criteria and scoring rubrics. Experimental analysis shows that our training regime strongly improves LLM-based scientific writing evaluation. Our models generalize effectively across tasks and to previously unseen scientific writing evaluation settings, allowing a single trained evaluator to be reused without task-specific retraining.

cs.CL↗

arXiv2026

Co-FactChecker: A Framework for Human-AI Collaborative Claim Verification Using Large Reasoning Models

Professional fact-checkers rely on domain knowledge and deep contextual understanding to verify claims. Large language models (LLMs) and large reasoning models (LRMs) lack such grounding and primarily reason from available evidence alone, creating a mismatch between expert-led and fully automated claim verification. To mitigate this gap, we posit human-AI collaboration as a more promising path forward, where expert feedback, grounded in real-world knowledge and domain expertise, guides the model's reasoning. However, existing LRMs are hard to calibrate to natural language feedback, particularly in a multi-turn interaction setup. We propose Co-FactChecker, a framework for human-AI collaborative claim verification. We introduce a new interaction paradigm that treats the model's thinking trace as a shared scratchpad. Co-FactChecker translates expert feedback into trace-edits that introduce targeted modifications to the trace, sidestepping the shortcomings of dialogue-based interaction. We provide theoretical results showing that trace-editing offers advantages over multi-turn dialogue, and our automatic evaluations demonstrate that Co-FactChecker outperforms existing autonomous and human-AI collaboration approaches. Human evaluations further show that Co-FactChecker is preferred over multi-turn dialogue, producing higher quality reasoning and verdicts along with relatively easier to interpret and more useful thinking traces.

cs.CL↗

arXiv2025

Illusion or Algorithm? Investigating Memorization, Emergence, and Symbolic Processing in In-Context Learning

Large-scale Transformer language models (LMs) trained solely on next-token prediction with web-scale data can solve a wide range of tasks after seeing just a few examples. The mechanism behind this capability, known as in-context learning (ICL), remains both controversial and poorly understood. Some studies argue that it is merely the result of memorizing vast amounts of data, while others contend that it reflects a fundamental, symbolic algorithmic development in LMs. In this work, we introduce a suite of investigative tasks and a novel method to systematically investigate ICL by leveraging the full Pythia scaling suite, including interim checkpoints that capture progressively larger amount of training data. By carefully exploring ICL performance on downstream tasks and simultaneously conducting a mechanistic analysis of the residual stream's subspace, we demonstrate that ICL extends beyond mere "memorization" of the training corpus, yet does not amount to the implementation of an independent symbolic algorithm. Our results also clarify several aspects of ICL, including the influence of training dynamics, model capabilities, and elements of mechanistic interpretability. Overall, our work advances the understanding of ICL and its implications, offering model developers insights into potential improvements and providing AI security practitioners with a basis for more informed guidelines.

cs.CL↗

arXiv2025

Problem Solving Through Human-AI Preference-Based Cooperation

While there is a widespread belief that artificial general intelligence (AGI) -- or even superhuman AI -- is imminent, complex problems in expert domains are far from being solved. We argue that such problems require human-AI cooperation and that the current state of the art in generative AI is unable to play the role of a reliable partner due to a multitude of shortcomings, including difficulty to keep track of a complex solution artifact (e.g., a software program), limited support for versatile human preference expression and lack of adapting to human preference in an interactive setting. To address these challenges, we propose HAICo2, a novel human-AI co-construction framework. We take first steps towards a formalization of HAICo2 and discuss the difficult open research problems that it faces.

cs.AI↗

arXiv2025

Factual Self-Awareness in Language Models: Representation, Robustness, and Scaling

Factual incorrectness in generated content is one of the primary concerns in ubiquitous deployment of large language models (LLMs). Prior findings suggest LLMs can (sometimes) detect factual incorrectness in their generated content (i.e., fact-checking post-generation). In this work, we provide evidence supporting the presence of LLMs' internal compass that dictate the correctness of factual recall at the time of generation. We demonstrate that for a given subject entity and a relation, LLMs internally encode linear features in the Transformer's residual stream that dictate whether it will be able to recall the correct attribute (that forms a valid entity-relation-attribute triplet). This self-awareness signal is robust to minor formatting variations. We investigate the effects of context perturbation via different example selection strategies. Scaling experiments across model sizes and training dynamics highlight that self-awareness emerges rapidly during training and peaks in intermediate layers. These findings uncover intrinsic self-monitoring capabilities within LLMs, contributing to their interpretability and reliability.

cs.CL↗

arXiv2024

Mechanistic Behavior Editing of Language Models

Large Language Models trained on web-scale text acquire language generation abilities that can solve a wide range of tasks, particularly when task knowledge is refined into the generative prior using in-context examples. However, spurious features learned from noisy data hinder their generalizability. Supervised finetuning can introduce task specificity, but introduce data inefficiency. Prior studies indicate that (i) noisy neural circuitries coexist with generalizable ones within LLMs, and (ii) finetuning typically enhances (or suppresses) existing abilities without introducing newer ones. Building upon these, we propose TaRot, a novel method for task adaptation. TaRot intervenes in the neural circuitries using learnable rotation matrices that are optimized using Bayesian Optimization, on labelled samples in the order of standard few-shot prompting examples. Experiments on multiple classification and generation tasks using LLMs of varying sizes reveal the efficacy of TaRot, improving upon both zero- as well as few-shot performance, with average improvements (across models and tasks) of 23.81% and 11.15%, respectively. The source code is available at https://github.com/joykirat18/TaRot

cs.CL↗

arXiv2024

Can LLMs replace Neil deGrasse Tyson? Evaluating the Reliability of LLMs as Science Communicators

Large Language Models (LLMs) and AI assistants driven by these models are experiencing exponential growth in usage among both expert and amateur users. In this work, we focus on evaluating the reliability of current LLMs as science communicators. Unlike existing benchmarks, our approach emphasizes assessing these models on scientific questionanswering tasks that require a nuanced understanding and awareness of answerability. We introduce a novel dataset, SCiPS-QA, comprising 742 Yes/No queries embedded in complex scientific concepts, along with a benchmarking suite that evaluates LLMs for correctness and consistency across various criteria. We benchmark three proprietary LLMs from the OpenAI GPT family and 13 open-access LLMs from the Meta Llama-2, Llama-3, and Mistral families. While most open-access models significantly underperform compared to GPT-4 Turbo, our experiments identify Llama-3-70B as a strong competitor, often surpassing GPT-4 Turbo in various evaluation aspects. We also find that even the GPT models exhibit a general incompetence in reliably verifying LLM responses. Moreover, we observe an alarming trend where human evaluators are deceived by incorrect responses from GPT-4 Turbo.

cs.CL↗

arXiv2024

Self-similarity of temporal interaction networks arises from hyperbolic geometry with time-varying curvature

The self-similarity of complex systems has been studied intensely across different domains due to its potential applications in system modeling, complexity analysis, etc., as well as for deep theoretical interest. Existing studies rely on scale transformations conceptualized over either a definite geometric structure of the system (very often realized as length-scale transformations) or purely temporal scale transformations. However, many physical and social systems are observed as temporal interactions among agents without any definitive geometry. Yet, one can imagine the existence of an underlying notion of distance as the interactions are mostly localized. Analysing only the time-scale transformations over such systems would uncover only a limited aspect of the complexity. In this work, we propose a novel technique of scale transformation that dissects temporal interaction networks under spatio-temporal scales, namely, flow scales. Upon experimenting with multiple social and biological interaction networks, we find that many of them possess a finite fractal dimension under flow-scale transformation. Finally, we relate the emergence of flow-scale self-similarity to the latent geometry of such networks. We observe strong evidence that justifies the assumption of an underlying, variable-curvature hyperbolic geometry that induces self-similarity of temporal interaction networks. Our work bears implications for modeling temporal interaction networks at different scales and uncovering their latent geometric structures.

physics.soc-ph↗

arXiv2024

Language Models can Exploit Cross-Task In-context Learning for Data-Scarce Novel Tasks

Large Language Models (LLMs) have transformed NLP with their remarkable In-context Learning (ICL) capabilities. Automated assistants based on LLMs are gaining popularity; however, adapting them to novel tasks is still challenging. While colossal models excel in zero-shot performance, their computational demands limit widespread use, and smaller language models struggle without context. This paper investigates whether LLMs can generalize from labeled examples of predefined tasks to novel tasks. Drawing inspiration from biological neurons and the mechanistic interpretation of the Transformer architecture, we explore the potential for information sharing across tasks. We design a cross-task prompting setup with three LLMs and show that LLMs achieve significant performance improvements despite no examples from the target task in the context. Cross-task prompting leads to a remarkable performance boost of 107% for LLaMA-2 7B, 18.6% for LLaMA-2 13B, and 3.2% for GPT 3.5 on average over zero-shot prompting, and performs comparable to standard in-context learning. The effectiveness of generating pseudo-labels for in-task examples is demonstrated, and our analyses reveal a strong correlation between the effect of cross-task examples and model activation similarities in source and target input tokens. This paper offers a first-of-its-kind exploration of LLMs' ability to solve novel tasks based on contextual signals from different task examples.

cs.CL↗

arXiv2024

How to think step-by-step: A mechanistic understanding of chain-of-thought reasoning

Despite superior reasoning prowess demonstrated by Large Language Models (LLMs) with Chain-of-Thought (CoT) prompting, a lack of understanding prevails around the internal mechanisms of the models that facilitate CoT generation. This work investigates the neural sub-structures within LLMs that manifest CoT reasoning from a mechanistic point of view. From an analysis of Llama-2 7B applied to multistep reasoning over fictional ontologies, we demonstrate that LLMs deploy multiple parallel pathways of answer generation for step-by-step reasoning. These parallel pathways provide sequential answers from the input question context as well as the generated CoT. We observe a functional rift in the middle layers of the LLM. Token representations in the initial half remain strongly biased towards the pretraining prior, with the in-context prior taking over in the later half. This internal phase shift manifests in different functional components: attention heads that write the answer token appear in the later half, attention heads that move information along ontological relationships appear in the initial half, and so on. To the best of our knowledge, this is the first attempt towards mechanistic investigation of CoT reasoning in LLMs.

cs.CL↗