INDEX

Explanations

phrases related to approval or authorization

oai_token-act-pair · gpt-3.5-turbo

phrases or terms related to official approvals or permissions

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

obs

-0.71

zanne

-0.66

ĸļ

-0.65

zilla

-0.62

onomic

-0.61

 Kahn

-0.61

jad

-0.61

ilant

-0.60

ISM

-0.59

Gall

-0.59

POSITIVE LOGITS

shoot

0.78

ments

0.71

undown

0.70

sites

0.68

abase

0.67

site

0.65

MENTS

0.65

ardon

0.63

loading

0.62

spr

0.62

Activations Density 0.041%

phrases related to approval or authorization

phrases or terms related to official approvals or permissions

No Comments

No Known Activations