INDEX

Explanations

references to European locations, especially those containing the prefix "eur."

oai_token-act-pair · gpt-3.5-turbo

mentions of European-related topics or entities

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

corn

-1.02

-0.66

 Lime

-0.64

 Chick

-0.64

 Huff

-0.64

Miller

-0.63

--------------------------------------------------------

-0.63

 Alabama

-0.62

smith

-0.61

 Solitaire

-0.61

POSITIVE LOGITS

ignty

0.92

onymous

0.84

 livest

0.84

 Myster

0.81

naire

0.80

izons

0.76

geoning

0.75

bley

0.74

nsic

0.74

mares

0.73

Activations Density 0.032%

references to European locations, especially those containing the prefix "eur."

mentions of European-related topics or entities

No Comments

No Known Activations