Goodfire's Silico Finds Human Cognitive Biases Geometrically Encoded Inside LLMs
Goodfire used mechanistic interpretability to find a 'bouba-kiki space' inside Llama and Gemma, and is opening access to Silico, the platform that found it.
- Goodfire found a "bouba-kiki space" inside Llama and Gemma: a geometric direction in model activations that separates spiky-sounding from round-sounding words, mirroring a century-old psycholinguistics result.
- This was found by probing internal activations directly, not by behavioral testing, showing the models absorbed cross-modal sound symbolism from training data.
- Silico, Goodfire's agentic interpretability platform, is now in private beta and can run end-to-end research experiments autonomously.
- Silico reproduced a months-long internal research project (RLFR) in two days, cutting hallucinations in Qwen3-8B by 37% with no capability drop.
- Goodfire has previously used interpretability to cut LLM hallucinations by 58% at 90x lower cost than LLM-as-judge methods.
- Silico is available by request with case-by-case pricing; Goodfire is a $1.25B company backed by Anthropic, Menlo Ventures, and others.
There's a well-known experiment in psycholinguistics: show someone two shapes, one spiky and one round, and ask which is called "kiki" and which is called "bouba." The bouba-kiki effect is a non-arbitrary mental association between certain speech sounds and certain visual shapes, and when presented with nonsense words, people tend to associate "soft" sounds like bouba with a rounded shape and "hard" sounds like kiki with a spiky shape. Now, Goodfire has found the same structure living inside the internal representations of large language models.
A direction in activation space that knows what "spiky" sounds like
Using their interpretability platform Silico, Goodfire's team probed the internal activations of Llama and Gemma and found a specific geometric direction in the models' hidden states that cleanly separates "spiky-sounding" words from "round-sounding" ones, independent of their actual meaning. Words like "kiki," "takete," and "crisp" cluster at one end; words like "bouba," "maluma," and "mellow" cluster at the other. This is mechanistic interpretability in action: instead of asking a model what it thinks, you look directly at the geometry of its internal representations.
The finding matters because it shows the models didn't just learn to answer bouba-kiki quiz questions correctly. They internalized the underlying structure of sound symbolism as a geometric feature of their representation space. The bouba-kiki effect has been robustly confirmed across a majority of cultures and languages, and its discovery dates back to the 1920s. The fact that LLMs replicate it geometrically suggests they absorbed this cross-modal regularity from the statistical patterns in their training text, not from any explicit instruction.