INDEX

Explanations

mentions of people or their actions in various contexts

oai_token-act-pair · gpt-3.5-turbo

repeated phrases indicating the presence of people and their actions or opinions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

rouse

-0.69

ð

-0.69

 entails

-0.63

ipient

-0.62

 Highlights

-0.61

izable

-0.61

ises

-0.60

 occurs

-0.60

izes

-0.60

 breach

-0.60

POSITIVE LOGITS

ateurs

0.99

 reluctant

0.99

 willing

0.96

 eager

0.95

wolves

0.93

 accustomed

0.93

 able

0.92

 afraid

0.91

 supposed

0.88

 aware

0.87

Activations Density 0.374%

mentions of people or their actions in various contexts

repeated phrases indicating the presence of people and their actions or opinions

No Comments

No Known Activations