INDEX

Explanations

proper nouns and abbreviations related to organizations, names, and titles that include the letter 'J'

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

avorite

-0.73

 sarc

-0.70

ãĤª

-0.68

pection

-0.67

 predec

-0.64

uity

-0.61

ishable

-0.60

theless

-0.60

uming

-0.60

 indispensable

-0.59

POSITIVE LOGITS

 Rowling

0.86

ansen

0.83

 Marriott

0.78

EStream

0.73

ernaut

0.73

dan

0.72

owicz

0.71

ournal

0.71

neys

0.71

usalem

0.71

Activations Density 0.353%

proper nouns and abbreviations related to organizations, names, and titles that include the letter 'J'

references to specific individuals and entities

No Comments

No Known Activations