28 accepted papers
Why do thinking language models outperform their base counterparts, and what exactly do they learn during training? We introduce constructive model diffing, a framework for understanding fine-tuned models by explicitly constructing the base-to-fine-tuned difference from interpretable components to p…
Recent studies indicate that when faced with explicit biases in prompts, models often omit mentioning these biases in their Chain-of-Thought (CoT) output, revealing that verbalized reasoning can give an incorrect picture of how models arrive at conclusions (unfaithfulness). In this work, we show tha…
Finetuning large language models on narrowly harmful datasets can cause them to become emergently misaligned, giving stereotypically `evil' responses across diverse unrelated settings. Concerningly, a pre-registered survey of experts failed to predict this result, highlighting our poor understanding…
Analyzing large-scale text corpora is a core challenge in machine learning, crucial for tasks like identifying undesirable model behaviors. Current methods often rely on costly LLM-based techniques (e.g. annotating dataset differences) or dense embedding models (e.g. for clustering), which lack cont…
Finetuning on narrow domains has become an essential tool to adapt Large Language Models (LLMs) to specific tasks and to create models with known unusual properties that are useful for safety research. Model diffing--the study of differences between base and finetuned models--is a promising approach…
Large language models (LLMs) can sometimes detect when they are being evaluated and adjust their behavior to appear more aligned, compromising the reliability of safety evaluations. In this paper, we show that adding a steering vector to an LLM's activations can suppress evaluation-awareness and mak…
Fine-tuning large language models (LLMs) can lead to unintended out-of-distribution generalization. Standard approaches to this problem rely on modifying the training data, for example by adding data that better specify the intended generalization. However, this is not always practical. We introduce…
We argue that interpreting reasoning models from a single chain-of-thought (CoT) is fundamentally inadequate. To understand computation and causal influence, one must study reasoning as a distribution of possible trajectories elicited by a given prompt. We approximate this distribution via on-policy…
Prior work suggests that language models, while trained on next token prediction, show implicit planning behavior: they may select the next token in preparation to a predicted future token, such as a likely rhyming word, as supported by a prior qualitative study of Claude 3.5 Haiku using a cross-lay…
Sparse autoencoders (SAEs) are a popular method for interpreting concepts represented in large language model (LLM) activations. However, there is a lack of evidence regarding the validity of their interpretations due to the lack of a ground truth for the concepts used by an LLM, and a growing numbe…
Hallucinations in large language models are a widespread problem, yet the mechanisms behind whether models will hallucinate are poorly understood, limiting our ability to solve this problem. Using sparse autoencoders as an interpretability tool, we discover that a key part of these mechanisms is ent…
Sparse Autoencoders (SAEs) are a popular method for decomposing Large Language Model (LLM) activations into interpretable latents, however they have a substantial training cost and SAEs learned on different models are not directly comparable. Motivated by relative representation similarity measures,…
Sparse autoencoders (SAEs) have emerged as a powerful tool for interpreting neural networks by extracting the concepts represented in their activations. However, choosing the size of the SAE dictionary (i.e. number of learned concepts) creates a tension: as dictionary size increases to capture more…
Model diffing is the study of how fine-tuning changes a model's representations and internal algorithms. Many behaviors of interest are introduced during fine-tuning, and model diffing offers a promising lens to interpret such behaviors. Crosscoders are a recent model diffing method that learns a…
Sparse autoencoders (SAEs) are a popular technique for interpreting language model activations, and there is extensive recent work on improving SAE effectiveness. However, most prior work evaluates progress using unsupervised proxy metrics with unclear practical relevance. We introduce SAEBench, a c…
Sparse autoencoders (SAEs) are a popular tool for interpreting large language model activations, but their utility in addressing open questions in interpretability remains unclear. In this work, we demonstrate their effectiveness by using SAEs to deepen our understanding of the mechanism behind in-c…
A common goal of mechanistic interpretability is to decompose the activations of neural networks into features: interpretable properties of the input computed by the model. Sparse autoencoders (SAEs) are a popular method for finding these features in LLMs, and it has been postulated that they can be…
Training vision language models (VLMs) aims to align visual representations from a vision encoder with the textual representations of a pretrained large language model (LLM). However, many VLMs exhibit reduced factual recall performance compared to their LLM backbones, raising the question of how ef…
Disentangling model activations into human-interpretable features is a central problem in interpretability. Sparse autoencoders (SAEs) have recently attracted much attention as a scalable unsupervised approach to this problem. However, our imprecise understanding of ground-truth features in realisti…
Despite their widespread use, the mechanisms by which large language models (LLMs) represent and regulate uncertainty in next-token predictions remain largely unexplored. This study investigates two critical components believed to influence this uncertainty: the recently discovered entropy neurons a…
Prior interpretability research studying narrow distributions has preliminarily identified self-repair, a phenomena where if components in large language models are ablated, later components will change their behavior to compensate. Our work builds off this past literature, demonstrating that self-r…
Recent work has found that sparse autoencoders (SAEs) are an effective technique for unsupervised discovery of interpretable features in language models' (LMs) activations, by finding sparse, linear reconstructions of those activations. We introduce the Gated Sparse Autoencoder (Gated SAE), which ac…
Mechanistic interpretability aims to attribute high-level model behaviors to specific, interpretable learned features. It is hypothesized that these features manifest as directions or low-dimensional subspaces within activation space. Accordingly, recent studies have explored the identification and…
Conversational large language models are fine-tuned for both instruction-following and safety, resulting in models that obey benign requests but refuse harmful ones. While this refusal behavior is widespread across chat models, its underlying mechanisms remain poorly understood. In this work, we sho…
Mechanistic interpretability seeks to understand the internal mechanisms of machine learning models, where localization—identifying the important model components—is a key step. Activation patching, also known as causal tracing or interchange intervention, is a standard technique for this task (Vig…
A key goal in mechanistic interpretability is circuit analysis: finding sparse subgraphs of models corresponding to specific behaviors or capabilities. However, MLP sublayers make fine-grained circuit analysis on transformer-based language models difficult. In particular, interpretable features—such…
Universality is a key hypothesis in mechanistic interpretability -- that different models learn similar features and circuits when trained on similar tasks. In this work, we study the universality hypothesis by examining how small networks learn to implement group compositions. We present a novel al…
Neural networks often exhibit emergent behavior in which qualitatively new capabilities that arise from scaling up the number of parameters, training data, or even the number of steps. One approach to understanding emergence is to find the continuous \textit{progress measures} that underlie the seem…