INDEX

Explanations

phrases related to instructions or descriptions, particularly those mentioning products or their features

oai_token-act-pair · gpt-3.5-turbo

elements of instructions or frameworks related to recipes and creative processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

-0.68

SpaceEngineers

-0.64

renheit

-0.62

 Username

-0.61

}}

-0.61

tion

-0.60

reenshots

-0.59

gae

-0.58

eanor

-0.57

 partName

-0.57

POSITIVE LOGITS

 that

1.23

that

1.20

 THAT

1.05

 whose

0.86

 which

0.85

 capable

0.83

 That

0.80

 thats

0.78

 conducive

0.77

 irresistible

0.77

Activations Density 0.690%

phrases related to instructions or descriptions, particularly those mentioning products or their features

elements of instructions or frameworks related to recipes and creative processes

No Comments

No Known Activations