INDEX

Explanations

phrases indicating a change in direction or action

oai_token-act-pair · gpt-3.5-turbo

instances of the word "That" indicating significant points or ideas within the text

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

hend

-0.71

aunder

-0.70

hus

-0.68

 Roses

-0.68

hat

-0.67

 Frames

-0.66

igmatic

-0.66

rypt

-0.65

igne

-0.65

feel

-0.64

POSITIVE LOGITS

 includes

1.11

 proposal

1.10

 entails

1.08

 involves

1.07

 effort

1.06

 announcement

1.01

 resulted

1.01

 contradicts

1.00

 requires

0.97

 eliminates

0.97

Activations Density 0.150%

phrases indicating a change in direction or action

instances of the word "That" indicating significant points or ideas within the text

No Comments

No Known Activations