INDEX

Explanations

mentions of the word "Fun" with varying activation strengths

oai_token-act-pair · gpt-3.5-turbo

mentions of "Fun" in various contexts related to entertainment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

xual

-0.74

Minimum

-0.70

ib

-0.69

×Ļ

-0.65

utical

-0.64

Øª

-0.64

txt

-0.63

ean

-0.62

BOOK

-0.61

ä»£

-0.60

POSITIVE LOGITS

nels

1.25

eral

1.22

nel

1.19

imation

1.10

nell

1.07

ctions

1.00

Fun

0.96

ilial

0.93

atorium

0.89

ctor

0.79

Activations Density 0.005%

mentions of the word "Fun" with varying activation strengths

mentions of "Fun" in various contexts related to entertainment

No Comments

No Known Activations