INDEX

Explanations

words related to specific geographical locations

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

rig

-0.66

 overhe

-0.66

 rigid

-0.64

Rig

-0.64

 Libre

-0.63

 resume

-0.62

 overflowing

-0.62

 locker

-0.61

 charger

-0.61

 Oper

-0.61

POSITIVE LOGITS

bush

4.85

vard

1.28

baugh

1.08

wang

1.05

Bush

1.02

nz

1.00

eus

1.00

baum

0.96

woods

0.96

 apiece

0.94

Activations Density 0.032%

words related to specific geographical locations

references to specific individuals or names

No Comments

No Known Activations