INDEX

Explanations

words or phrases that are often labeled or categorized as something specific (often with the term "so-called")

oai_token-act-pair · gpt-3.5-turbo

the phrase "so-called" as an introduction to a term or label

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

erves

-0.77

=-=-=-=-

-0.71

 nevertheless

-0.70

 furthermore

-0.66

rehens

-0.66

ÃŃs

-0.65

-0.64

,,,,,,,,

-0.62

ATCH

-0.62

EEE

-0.62

POSITIVE LOGITS

"#

0.89

judicial

0.87

metadata

0.82

ordinary

0.81

0.80

social

0.78

 caliphate

0.77

advant

0.76

 Islamic

0.75

prime

0.69

Activations Density 0.036%

words or phrases that are often labeled or categorized as something specific (often with the term "so-called")

the phrase "so-called" as an introduction to a term or label

No Comments

No Known Activations