INDEX

Explanations

numeric values related to quantities or figures specified in a document

oai_token-act-pair · gpt-3.5-turbo

numerical values and phrases related to legislative processes

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

rome

-0.81

 metab

-0.70

 Syndrome

-0.69

 Thumbnails

-0.66

ĸļ

-0.65

 circa

-0.65

 Sunder

-0.64

 Mania

-0.64

ixtape

-0.62

 Scourge

-0.61

POSITIVE LOGITS

 senators

1.25

 lawmakers

1.17

 justices

1.16

 legislators

1.12

 signatures

1.10

 amendments

1.04

 commissioners

1.01

 states

0.98

 votes

0.97

 Senators

0.96

Activations Density 0.163%

numeric values related to quantities or figures specified in a document

numerical values and phrases related to legislative processes

No Comments

No Known Activations