INDEX

Explanations

places or buildings where organizations are located or operate from

oai_token-act-pair · gpt-3.5-turbo

instances of the word "headquarters."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

erer

-0.83

spir

-0.80

theless

-0.74

flix

-0.74

pedia

-0.74

ef

-0.73

nel

-0.73

erers

-0.70

isher

-0.69

ulus

-0.68

POSITIVE LOGITS

 located

0.91

 headquarters

0.91

 building

0.90

 buildings

0.83

enture

0.80

 premises

0.80

 Building

0.78

 location

0.74

 address

0.74

 annex

0.73

Activations Density 0.045%

places or buildings where organizations are located or operate from

instances of the word "headquarters."

No Comments

No Known Activations