INDEX

Explanations

phrases related to new methods or opportunities for innovation

oai_token-act-pair · gpt-3.5-turbo

phrases indicating personal experiences or new opportunities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

rake

-0.58

 arrang

-0.56

ammy

-0.55

 issuance

-0.55

 Ammo

-0.54

 scapego

-0.51

urated

-0.51

isbury

-0.51

 sided

-0.51

 appointing

-0.50

POSITIVE LOGITS

 otherwise

1.20

 hitherto

1.15

 previously

1.09

 sorely

1.07

 unheard

1.03

 seldom

1.02

 hadn

1.02

 unimaginable

1.01

 never

1.00

 scarcely

0.97

Activations Density 0.320%

phrases related to new methods or opportunities for innovation

phrases indicating personal experiences or new opportunities

No Comments

No Known Activations