INDEX

Explanations

references to temporary or interim positions, specifically those involving legal or government roles

oai_token-act-pair · gpt-3.5-turbo

instances of the word "acting" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

xual

-0.85

joy

-0.77

pled

-0.73

front

-0.71

psy

-0.70

origin

-0.67

tch

-0.67

lake

-0.66

nda

-0.66

ijn

-0.65

POSITIVE LOGITS

uary

0.88

uate

0.78

 verbs

0.73

ional

0.73

 acting

0.70

uated

0.69

iates

0.69

uations

0.68

 overtime

0.67

inic

0.67

Activations Density 0.021%

references to temporary or interim positions, specifically those involving legal or government roles

instances of the word "acting" in various contexts

No Comments

No Known Activations