INDEX

Explanations

phrases indicating a transition or continuation in a speech or text

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a preface or introduction to content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

cow

-0.81

knit

-0.78

rod

-0.76

pless

-0.70

scl

-0.66

pee

-0.66

 Tens

-0.65

nian

-0.65

tek

-0.65

pak

-0.65

POSITIVE LOGITS

 doubt

1.30

ado

1.29

 hesitation

1.25

 provocation

1.08

 exaggeration

1.06

 explanation

1.02

 delay

1.00

 contradiction

0.92

 qualification

0.92

 irony

0.90

Activations Density 0.107%

phrases indicating a transition or continuation in a speech or text

phrases indicating a preface or introduction to content

No Comments

No Known Activations