INDEX

Explanations

the word "Free" followed by a variety of different terms

oai_token-act-pair · gpt-3.5-turbo

references to the term "Free."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

acea

-0.68

IOR

-0.63

 bore

-0.62

stroke

-0.62

essa

-0.61

 aggrav

-0.60

isks

-0.60

TIT

-0.59

 leth

-0.58

ihu

-0.57

POSITIVE LOGITS

 Free

3.89

Free

2.84

 free

2.29

free

2.25

 FREE

2.15

FREE

1.91

 Freedom

1.26

 Freed

1.12

Freedom

1.11

 Freeze

1.11

Activations Density 0.015%

the word "Free" followed by a variety of different terms

references to the term "Free."

No Comments

No Known Activations