INDEX

Explanations

specific mentions of the word 'type'

oai_token-act-pair · gpt-3.5-turbo

references to types and classifications in design and technology contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Leaks

-0.68

 Bills

-0.67

ahar

-0.65

 Bots

-0.61

 Appropriations

-0.61

eor

-0.60

Roz

-0.59

 Morning

-0.58

iae

-0.58

 Wake

-0.57

POSITIVE LOGITS

face

1.45

faces

1.38

ahead

1.15

casting

1.14

etting

1.11

etter

1.05

cast

1.02

alias

0.92

 locality

0.91

 certific

0.91

Activations Density 0.054%

specific mentions of the word 'type'

references to types and classifications in design and technology contexts

No Comments

No Known Activations