INDEX

Explanations

words related to causing action or response

oai_token-act-pair · gpt-3.5-turbo

phrases or words indicating triggers for actions or responses

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

atum

-0.85

çĦ

-0.79

wn

-0.75

oyal

-0.73

nect

-0.72

Í

-0.71

framework

-0.71

thin

-0.70

ãĥİ

-0.70

illusion

-0.69

POSITIVE LOGITS

 outcry

1.12

 inquiries

1.00

 warnings

0.98

 widespread

0.96

 speculation

0.96

 complaints

0.95

 questions

0.92

 accusations

0.91

 cries

0.91

 condemnation

0.91

Activations Density 0.074%

words related to causing action or response

phrases or words indicating triggers for actions or responses

No Comments

No Known Activations