question

active

question:do-llms-have-a-unified-representation-of-truth-that-spans-structurally-and-topically-diverse-data

Do LLMs have a unified representation of truth that spans structurally and topically diverse data?

Central research question driving dataset design and experimental approach

Source paper

extracted_from

The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets

(2023) · Samuel Marks · Max Tegmark

Neighborhood — ranked by edge-count

Papers (1)

paper

The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
associated_with

Findings (1)

finding

For LLaMA-2-70B, probes trained on larger_than+smaller_than achieve >95% accuracy on sp_en_trans regardless of probing technique
answered_by
Striking cross-domain generalization result supporting the claim that larger models represent abstract truth

Claims (1)

claim

LLMs linearly represent truth-relevant information beyond the plausibility of text, as evidenced by probes trained on likely performing poorly on anti-correlated datasets
gates
Establishes that the observed linear structure is not merely a representation of text probability

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

As LLMs scale, they develop increasingly general abstractions, with large models linearly representing abstract concepts like truth that capture shared properties of diverse inputsclaim0.839
Interpretive claim connecting scale to abstraction level in LLM representations
Truth may be linearly separable in the model's representation space, but the structure is richer than a single linear axisclaim0.821
Interpretive synthesis of DIM and cone intervention successes
In intermediate regimes of scale or layer depth, LLMs may linearly represent features at intermediate levels of abstraction such as 'accurate factual recall' or 'close association' rather than abstract truthclaim0.817
Theoretical interpretation of antipodal alignment and misalignment phenomena in PCA visualizations
LLM representations exhibit intriguing patterns under spatio-permutational analyses, suggesting a potentially profound yet tentative indication of consciousness.claim0.805
Qualified positive claim from spatio permutation analysis where two cases satisfy all three criteria.
Current LLMs cannot faithfully represent transformative experiences with epistemically opaque outcomes.claim0.802
Linear truth directions in LLMs are reliable primarily in factual recall cases and break down when truth assessment depends on computing and storing intermediate results.claim0.801
Central empirical conclusion of the paper about the fundamental limits of truth directions.
We hypothesize that LLMs represent correctness of arithmetic expressions differently from factual statements.hypothesis0.796
Core working hypothesis motivating the factual vs. arithmetic task split in the experimental design.
Whether conclusions about latent reflection directions generalize to larger LLMs, different architectures, or broader datasets remains to be verified.question0.792
Key limitation and open question about experimental scope.