INDEX

Explanations

words related to trickery or clever deception

oai_token-act-pair · gpt-3.5-turbo

references to clever tactics or techniques

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Domain

-0.67

Found

-0.65

Ide

-0.63

çĦ

-0.63

 isot

-0.61

 concess

-0.61

BW

-0.60

 Predators

-0.59

 Expend

-0.58

ãĥĺãĥ©

-0.57

POSITIVE LOGITS

ery

1.53

ster

1.37

sters

1.29

eries

1.21

iest

1.07

door

1.01

les

1.01

ett

0.94

 tricks

0.93

ety

0.93

Activations Density 0.048%

words related to trickery or clever deception

references to clever tactics or techniques

No Comments

No Known Activations