INDEX

Explanations

cities or locations

oai_token-act-pair · gpt-3.5-turbo

specific geographical locations and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Britann

-0.62

dule

-0.61

cession

-0.60

 Bleach

-0.59

zig

-0.58

orer

-0.57

Redditor

-0.57

plane

-0.57

unks

-0.56

mobi

-0.56

POSITIVE LOGITS

VILLE

1.13

ARY

1.09

OND

1.07

ION

1.03

IAL

1.03

YN

1.02

IAN

1.02

ENN

1.01

LAND

1.01

WARD

1.00

Activations Density 0.066%

cities or locations

specific geographical locations and organizations

No Comments

No Known Activations