INDEX

Explanations

mentions of the term "American" or related phrases

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "American" or related themes

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

order

-0.78

fill

-0.76

fm

-0.74

Fill

-0.70

orders

-0.69

lass

-0.68

theless

-0.68

ault

-0.68

Els

-0.68

SHARE

-0.67

POSITIVE LOGITS

 embassy

0.97

 Embassy

0.96

 Airlines

0.95

 diplomats

0.90

 Samoa

0.90

 embassies

0.88

AAF

0.86

 Expedition

0.83

 spies

0.83

 Ambassador

0.82

Activations Density 0.052%

mentions of the term "American" or related phrases

references to the concept of "American" or related themes

No Comments

No Known Activations