INDEX

Explanations

interview-related terms or mentions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "interview" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.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

math

-0.77

axy

-0.76

mil

-0.71

ciplinary

-0.71

blue

-0.70

 Goods

-0.68

 rejoice

-0.67

 Haram

-0.65

DIV

-0.65

bane

-0.63

POSITIVE LOGITS

ees

1.18

 interviews

0.95

 interview

0.93

 transcripts

0.91

ee

0.88

 Transcript

0.87

 subjects

0.86

 conducted

0.83

ioned

0.80

 Interview

0.79

Activations Density 0.045%

interview-related terms or mentions

instances of the word "interview" and its variations

No Comments

No Known Activations