INDEX

Explanations

words related to an intermediate level of expertise or a state in between two extremes

oai_token-act-pair · gpt-3.5-turbo

references to intermediate or transitional concepts

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

oran

-0.79

 Kinnikuman

-0.78

advertising

-0.69

ILLE

-0.66

ore

-0.65

©¶æ

-0.64

atche

-0.62

hold

-0.60

Paw

-0.60

âķĲâķĲ

-0.60

POSITIVE LOGITS

 intermediate

1.10

 intermedi

0.90

essor

0.83

yip

0.80

igmatic

0.79

 Intermediate

0.79

rador

0.76

yout

0.75

yrinth

0.75

clair

0.74

Activations Density 0.010%

words related to an intermediate level of expertise or a state in between two extremes

references to intermediate or transitional concepts

No Comments

No Known Activations