INDEX

Explanations

locations associated with government or political activities, specifically focusing on the Oval Office

oai_token-act-pair · gpt-3.5-turbo

mentions of the "Oval Office."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

terday

-0.74

weeney

-0.72

ifiable

-0.71

CVE

-0.68

ailing

-0.66

ãĤ½

-0.66

unks

-0.65

icity

-0.65

ayer

-0.64

 Survive

-0.64

POSITIVE LOGITS

 Office

0.96

croft

0.85

 Oval

0.82

 Depot

0.78

robe

0.76

 Lawn

0.75

 Room

0.75

office

0.74

 office

0.73

stad

0.72

Activations Density 0.028%

locations associated with government or political activities, specifically focusing on the Oval Office

mentions of the "Oval Office."

No Comments

No Known Activations