INDEX

Explanations

words related to deception and disguises

oai_token-act-pair · gpt-3.5-turbo

references to illusions and disguises

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

uled

-0.72

aldi

-0.69

orough

-0.69

Ü

-0.66

capacity

-0.65

vez

-0.64

acid

-0.63

 Issues

-0.63

olved

-0.62

iaries

-0.62

POSITIVE LOGITS

 deceive

1.10

 disgu

1.00

mir

0.94

 illusion

0.90

 deception

0.85

 disguise

0.84

 camoufl

0.84

pas

0.82

querade

0.81

 Illusion

0.80

Activations Density 0.070%

words related to deception and disguises

references to illusions and disguises

No Comments

No Known Activations