INDEX

Explanations

job-related information, such as hiring, firing, salaries, positions, and work schedules

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

powder

-0.79

Ô

-0.73

 Juda

-0.72

hran

-0.72

 Persia

-0.70

itsch

-0.68

pox

-0.68

opy

-0.66

 legalize

-0.64

rencies

-0.64

POSITIVE LOGITS

 jobs

1.42

 employees

1.36

job

1.35

 workforce

1.35

 employee

1.32

employ

1.29

Employ

1.28

 employment

1.28

 Employees

1.27

Job

1.25

Activations Density 0.898%

job-related information, such as hiring, firing, salaries, positions, and work schedules

No Comments

No Known Activations