INDEX

Explanations

verbs indicating causation or driving force

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate causation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

raq

-0.80

naissance

-0.77

kered

-0.75

asy

-0.75

SPONSORED

-0.73

 resil

-0.72

icket

-0.72

itled

-0.71

hement

-0.71

apest

-0.70

POSITIVE LOGITS

 virtue

1.32

products

1.05

laws

0.92

 fiat

0.89

 multiplying

0.79

 afar

0.77

 means

0.75

 leaps

0.74

 default

0.71

 STATS

0.70

Activations Density 0.272%

verbs indicating causation or driving force

phrases that indicate causation

No Comments

No Known Activations