INDEX

Explanations

keywords related to flags or banners

oai_token-act-pair · gpt-3.5-turbo

the term 'gags' and its variations in different contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 streng

-0.80

URRENT

-0.67

 carbohyd

-0.64

 mutual

-0.62

oun

-0.62

ibur

-0.62

 Seym

-0.61

 reconstruction

-0.60

 predec

-0.59

 Edmund

-0.59

POSITIVE LOGITS

glers

1.02

ags

0.98

gers

0.96

ging

0.90

zag

0.87

ration

0.87

ged

0.85

hi

0.83

gery

0.83

eries

0.81

Activations Density 0.007%

keywords related to flags or banners

the term 'gags' and its variations in different contexts

No Comments

No Known Activations