INDEX

Explanations

mentions of superlatives or extreme characteristics, such as 'largest' or 'most'

oai_token-act-pair · gpt-3.5-turbo

references to "the" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

BIL

-0.75

oult

-0.74

wards

-0.74

BILITIES

-0.73

isson

-0.70

ettings

-0.67

atz

-0.65

uitous

-0.64

Ã¨re

-0.64

BILITY

-0.63

POSITIVE LOGITS

 world

1.54

 country

1.33

 entire

1.29

 nation

1.27

 universe

1.19

 history

1.13

 WORLD

1.08

 globe

1.08

 whole

1.07

 shortest

1.04

Activations Density 0.097%

mentions of superlatives or extreme characteristics, such as 'largest' or 'most'

references to "the" in various contexts

No Comments

No Known Activations