INDEX

Explanations

the word "face" or words derived from it

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "fac" or its derivatives, typically related to facial features or aspects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Slug

-0.77

 Breed

-0.73

 Seas

-0.71

 Barkley

-0.68

 Sawyer

-0.68

UCK

-0.68

 grandfather

-0.67

 Polar

-0.67

 Salvation

-0.67

 Passage

-0.66

POSITIVE LOGITS

ilitation

1.28

ulty

1.09

eless

1.04

ilit

1.03

ilities

1.00

ilitating

0.96

ility

0.95

fac

0.93

sim

0.92

ework

0.91

Activations Density 0.018%

the word "face" or words derived from it

occurrences of the term "fac" or its derivatives, typically related to facial features or aspects

No Comments

No Known Activations