Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads

Aryo Pradipta Gema, Beatrice Alex, Pasquale Minervini

Keywords: Methods (probing, steering, causal interventions)

Other Keywords: Attention Heads, OV Circuit

TL;DR: We propose Logit-Contribution Scoring (LOCOS) which scores attention heads by projecting what they write at each source position onto the answer-token unembedding direction.

Abstract: In long-context use, large language models frequently synthesize answers from the meaning of a relevant context span rather than literally copy-pasting them. Identifying which attention heads perform this synthesis matters for interpreting long-context model behavior. Yet existing detectors miss these heads by construction: they reward heads whose attended token matches the generated token, a literal-copy criterion that captures where a head reads but not what it writes through its output-value (OV) circuit, the very mechanism that carries non-literal retrieval. We introduce Logit-Contribution Scoring (LOCOS), a write-aware detector that scores each head by the projection of its OV-circuit output onto the answer-token unembedding direction, contrasting needle and off-needle source positions in a single forward pass. Across three model families (Qwen3, Gemma-3, OLMo-3.1), mean-ablating the top LOCOS heads on the NoLiMa non-literal retrieval benchmark collapses ROUGE-L at lower head counts than every attention-based baseline; on Qwen3-8B, ablating 50 heads drives ROUGE-L from 0.401 to 0.000 while the strongest baseline still retains 0.292. The selected heads are retrieval-specific: parametric recall and arithmetic reasoning stay at baseline under the same ablation. On Qwen3-8B, the same ablation also drops MuSiQue from 0.55 to 0.08 and BABILong from 0.62 to 0.20, while a random-heads control stays within 0.05 of baseline.

Submission Number: 183