INDEX

Explanations

proper nouns or names led by someone or some group

oai_token-act-pair · gpt-3.5-turbo

phrases indicating leadership or a guiding role 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 @ 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

ypes

-0.67

 DISTRICT

-0.60

Effective

-0.58

 competitiveness

-0.56

 Vaugh

-0.56

 Address

-0.55

 warmth

-0.55

 Scotia

-0.54

 thirds

-0.54

 blades

-0.54

POSITIVE LOGITS

gers

1.22

ges

1.11

ger

1.11

gling

1.04

better

1.03

ging

1.02

ged

0.86

gest

0.85

dit

0.82

*/(

0.81

Activations Density 0.043%

proper nouns or names led by someone or some group

phrases indicating leadership or a guiding role in various contexts

No Comments

No Known Activations