INDEX

Explanations

locations or places mentioned in texts

oai_token-act-pair · gpt-3.5-turbo

locations and specific addresses or coordinates

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

METHOD

-0.69

sticks

-0.69

film

-0.69

tools

-0.69

hops

-0.65

cloth

-0.65

HTML

-0.64

hop

-0.63

biased

-0.63

henko

-0.62

POSITIVE LOGITS

rium

1.31

las

0.97

 dusk

0.95

 least

0.93

 Mile

0.87

abase

0.85

yp

0.84

mosp

0.83

mega

0.82

 midday

0.81

Activations Density 0.109%

locations or places mentioned in texts

locations and specific addresses or coordinates

No Comments

No Known Activations