Post-training effects on LLM vocabulary
Parent: Why LLMs converge on the same words: mode collapse and post-training lexical effects · Published reference · snapshot 2026-09-23
↓ Facts as markdownall context files
Source-anchored research for Post-training effects on LLM vocabulary. Run status: COMPLETED.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
Measurements and reference values
- The popular explanation that 'delve' is overused because OpenAI's outsourced RLHF annotators in Nigeria favor Nigerian formal English was proposed by Alex Hern and is framed as a plausible theory; circumstantial support is PubMed 'delve' usage rising 10-100x, with no direct causal evidence. (medium confidence) [source]
Problems, failure modes and limitations
- Current lexical-shift studies cannot test ChatGPT directly because it lacks a public base variant, and they use mid-size open models on English scientific text, limiting generalization to frontier chat assistants. (low confidence) [source]
Comparisons and alternatives
- Top overused lemmas in instruction-tuned models versus human abstracts include suggest, highlight, enhance, furthermore, additionally, crucial, potential, strategy; many are ordinary scientific vocabulary, so models amplify already-typical wording rather than inventing novel words. (low confidence) [source]
- Base models also diverge from human word usage (misalignment scores 0.0329-0.0345 vs human baseline 0.0304), so not all lexical drift is post-training; content-word alignment trends are weakened or reversed after instruction tuning. (low confidence) [source]
Facts and statements
- A 2026 study of 42,000 PubMed abstract continuations across six open model families (Falcon, Gemma, Llama, Mistral, OLMo, Yi; 4B-8B) defines curation-free metrics (Lexical Alignment Score, Triangulated Preference Shift) to isolate overuse relative to human text and shifts attributable to preference-learning stages. (medium confidence) [source]
- Instruction tuning does not uniformly change lexical diversity: in Qwen2.5-7B, Mistral-7B and Llama-3.1-8B on three multiple-choice benchmarks, cross-rationale similarity diversity (1-SelfBLEU) consistently fell while surface Unique-2 diversity moved in both directions, so 'post-training narrows vocabulary' is model- and metric-dependent. (medium confidence) [source]
- Markdown-heavy training data leaves stylistic fingerprints such as em-dash frequency; post-training can modify but does not fully erase these markers. (low confidence) [source]
Children
- No children recorded.