INDEX

Explanations

names of people or places

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and titles

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

 supers

-0.68

 shuttle

-0.68

COM

-0.67

TRA

-0.66

 stray

-0.66

 operating

-0.65

 roadside

-0.64

 metic

-0.64

 ecstasy

-0.63

 hemor

-0.63

POSITIVE LOGITS

ason

1.08

chuk

1.02

wald

1.01

ault

1.00

orst

1.00

beck

0.98

reau

0.97

ouf

0.97

iak

0.97

aga

0.96

Activations Density 0.170%

names of people or places

proper nouns, particularly names and titles

No Comments

No Known Activations