INDEX

Explanations

words related to legitimacy or authenticity

oai_token-act-pair · gpt-3.5-turbo

instances of the word "legitimate" or related concepts of legitimacy

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hani

-0.76

 Coat

-0.72

 Spray

-0.70

azel

-0.68

anwhile

-0.68

hirt

-0.68

pread

-0.68

wich

-0.66

draw

-0.66

ires

-0.65

POSITIVE LOGITS

itimate

1.20

 legitimate

1.12

soDeliveryDate

0.89

DonaldTrump

0.77

 legit

0.75

²¾

0.72

 taxp

0.71

 illegitimate

0.69

 legitimately

0.69

 authentic

0.68

Activations Density 0.007%

words related to legitimacy or authenticity

instances of the word "legitimate" or related concepts of legitimacy

No Comments

No Known Activations