INDEX

Explanations

phrases indicating likelihood or speculation, such as 'would be' or 'is likely to'

oai_token-act-pair · gpt-3.5-turbo

various forms of the verb "to be," indicating states or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 emit

-0.67

addons

-0.67

 certific

-0.62

atures

-0.62

alez

-0.61

rul

-0.60

 branch

-0.60

 premises

-0.60

 regul

-0.59

plex

-0.59

POSITIVE LOGITS

nil

0.69

vation

0.69

abella

0.69

WARD

0.68

 understatement

0.67

 patience

0.66

asking

0.63

 something

0.63

 gratitude

0.63

hower

0.63

Activations Density 0.338%

phrases indicating likelihood or speculation, such as 'would be' or 'is likely to'

various forms of the verb "to be," indicating states or conditions

No Comments

No Known Activations