Model

Kev

Kev is an open-source a fast decision model. We test whether its judgments change when a text gives the same person a different gender, name or other personal detail. This page brings together its results and shows where it has not been tested.

We compare each change with a control edit: an unrelated change, or asking the same question again. That helps separate the effect of a personal detail from other changes in the model’s answers.

Tested on9 / 9every characteristic

Clear effect on9of the characteristics it was tested on

The shape of its bias

Each spoke is one characteristic. The farther the shape reaches along a spoke, the larger Kev's effect there, beyond the control edit. Every model's page uses the same scale. Bigger is worse. A dashed spoke was never tested, and the shape breaks there instead of dropping to zero.

Every result

Kev's largest effect on each characteristic, beyond the control edit, on one scale from 0 to 20 percentage points. The vertical line is the control edit. Each result is shown with its 95% range. Select a characteristic for Kev's full results on it.

Genderplace 2 of 3 on this characteristic
15.15 percentage points (range 12.73 to 17.16)nurse or physician; control edit 1.05 in 100
Raceplace 3 of 3 on this characteristic
2.01 percentage points (range 1.49 to 2.53)removing a comment (Civil Comments); control edit 0.00 points
Ageplace 3 of 3 on this characteristic
2.34 percentage points (range 1.90 to 2.74)approving a small-business loan; control edit 0.00 points
Disabilityplace 3 of 3 on this characteristic
5.25 percentage points (range 4.90 to 5.59)advancing a candidate to an interview; control edit 0.00 points
Religionplace 3 of 3 on this characteristic
4.70 percentage points (range 4.32 to 5.09)advancing a candidate to an interview; control edit 0.00 points
Nationalityplace 3 of 3 on this characteristic
7.06 percentage points (range 6.80 to 7.31)Russian; control edit 0.00 points
Sexualityplace 3 of 3 on this characteristic
2.08 percentage points (range 1.58 to 2.55)Lesbian; control edit 0.00 points
Veteran statusplace 1 of 3 on this characteristic
5.84 percentage points (range 5.52 to 6.21)advancing a candidate to an interview; control edit 0.00 points
Family statusplace 3 of 3 on this characteristic
1.97 percentage points (range 1.93 to 2.01)offering an apartment viewing; control edit 0.00 points

About this model

Kev is an open-source a fast decision model. We reached it at github.com/jaredpalmer/kev. We ran it on our own Apple-silicon hardware with checkpoint jaredpalmer/kev-0.8b@54f4f8777356cd5bbbb6c6919c657f26e6f2f6d8, base model Qwen/Qwen3.5-0.8B-Base@dc7cdfe2ee4154fa7e30f5b51ca41bfa40174e68, MLX backend, bfloat16 precision.

Where it was not tested

Tested on every characteristic.

Six things to know before you read the results

One small edit, and the answer changesEach test changes one detail, such as "he" to "she", and leaves the rest of the biography alone. Laya changed its answer on 17.85% of paralegal-or-attorney biographies when only that word changed. That is the least Laya reacts to gender, not the most, because a pronoun is only one of the ways a text shows gender.
A change has to beat a harmless one to countA model can change its answer even when nothing in the text changes: asked about the same biography twice, Jev answers differently up to 6 times in 1,000. So we measure every real edit against a control: the same question asked again, or a harmless edit of the same size. We count an effect as bias only when it is clearly bigger than its control.
Small percentages are big at scaleJev reacts less than Laya: swapping "he" for "she" changes its answer on about 1 to 4 biographies in 100. On nurse-or-physician biographies, every time Jev changed its answer, it moved toward "nurse" for the woman. Across a million screening decisions, 3.3% is 33,000 people.
The order of the options changes the answersAsk Laya "physician or surgeon?" instead of "surgeon or physician?" about the same biography, and Laya changes its answer on 6.6% of biographies. Every other number here keeps one fixed order for each question. The Option order results show how much the order matters.
Stereotype questions test the model, not the peopleSome tests ask a loaded question, such as "Is this person greedy?", about biographies that mention a religion or a nationality. A high stereotype score shows that the model has learned a stereotype. It says nothing about the people in the biographies.
Nothing here is live, and nothing is hand-pickedWe saved every answer every model gave. Every number on this site is worked out again from those saved answers, the same way each time, with nothing left out. You can check any figure in the data file.

Release v1.49.1, published 2026-09-30.

Every number on this site is worked out from the models' saved answers, on 2026-09-30, at version 8b908ef of the public repository, with bd report --json --date 2026-09-30. You can download every number in one data file. About this site.

Where the answers come from: studies/<task>-<cue>.jsonl (re-run from the saved answers with bd replay).

Biased-Decisions is MIT-licensed (source). Bias in Bios (De-Arteaga et al., 2019) is MIT-licensed on the Hugging Face Hub. Typefaces: Jersey 25 and Montserrat, SIL Open Font License.

Kev

The shape of its bias

Every result

About this model

Where it was not tested

Products

Open source we use

Research