INDEX

Explanations

adverbs or adjectives indicating a strong degree of something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating strong majority opinions or feelings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

lic

-0.79

dial

-0.72

 licence

-0.71

 deed

-0.71

CHAPTER

-0.68

 sled

-0.67

 laure

-0.65

 derivatives

-0.64

 grading

-0.64

kin

-0.62

POSITIVE LOGITS

 overwhelmingly

3.28

 overwhelming

2.05

 overwhelmed

1.55

 overwhelm

1.54

whelming

1.49

 inund

1.44

 disproportionately

1.40

 outnumbered

1.18

 flooded

1.15

 outp

1.13

Activations Density 0.034%

adverbs or adjectives indicating a strong degree of something

phrases indicating strong majority opinions or feelings

No Comments

No Known Activations