INDEX

Explanations

instances of a positive attribute being described in superlatives

oai_token-act-pair · gpt-3.5-turbo

references to the word "nic" and its variations used in different contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

WIND

-0.84

 Continued

-0.73

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.70

 Vanity

-0.69

é¾

-0.69

ëĭ

-0.67

 Shogun

-0.66

Stage

-0.66

è¦ļéĨĴ

-0.66

 Sorceress

-0.65

POSITIVE LOGITS

eties

1.50

nic

1.12

otin

1.10

uity

1.06

ety

1.01

opher

1.01

uously

1.00

otine

0.98

esse

0.97

olas

0.97

Activations Density 0.010%

instances of a positive attribute being described in superlatives

references to the word "nic" and its variations used in different contexts

No Comments

No Known Activations