INDEX

Explanations

names of specific cities or locations

oai_token-act-pair · gpt-3.5-turbo

mentions of the city of Frankfurt

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

orthy

-0.77

enegger

-0.74

estern

-0.74

arsity

-0.73

ersive

-0.72

cknowled

-0.71

earance

-0.70

chairs

-0.69

FUL

-0.65

earing

-0.64

POSITIVE LOGITS

furt

1.15

 Frankfurt

1.02

wich

0.84

 Lumpur

0.78

enburg

0.74

enberg

0.70

keley

0.69

 Airport

0.68

enstein

0.68

ford

0.68

Activations Density 0.008%

names of specific cities or locations

mentions of the city of Frankfurt

No Comments

No Known Activations