INDEX

Explanations

words related to disguises and deception

oai_token-act-pair · gpt-3.5-turbo

variations of the word "guise" indicating deception or disguise

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

hower

-0.84

ŃĶ

-0.75

 Spectrum

-0.74

cling

-0.72

 spect

-0.72

 backdrop

-0.67

croft

-0.65

è¦ļéĨĴ

-0.64

 cycle

-0.63

cycle

-0.63

POSITIVE LOGITS

arant

1.13

idelines

1.10

inea

1.08

arding

1.06

vernment

1.04

ilty

1.03

errilla

1.01

pta

0.98

ests

0.96

atem

0.95

Activations Density 0.013%

words related to disguises and deception

variations of the word "guise" indicating deception or disguise

No Comments

No Known Activations