INDEX

Explanations

phrases related to workforce changes, especially layoff-related language

oai_token-act-pair · gpt-3.5-turbo

mentions of layoffs and job reductions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.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

netflix

-0.87

 prosecutions

-0.77

aging

-0.76

ability

-0.72

ometers

-0.66

ety

-0.66

Lib

-0.66

 Adin

-0.64

activated

-0.64

 nudity

-0.63

POSITIVE LOGITS

OULD

0.91

 groundwork

0.87

OOL

0.86

 layoffs

0.85

OHN

0.83

 WARN

0.73

ffen

0.71

ABE

0.69

IAL

0.68

iesel

0.67

Activations Density 0.030%

phrases related to workforce changes, especially layoff-related language

mentions of layoffs and job reductions

No Comments

No Known Activations