INDEX

Explanations

legal and political terms, including words related to ethics and institutions

oai_token-act-pair · gpt-3.5-turbo

phrases related to hate speech and its implications

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

)",

-0.79

[|

-0.75

iHUD

-0.73

\">

-0.73

 analyse

-0.70

 mixes

-0.69

 forecast

-0.68

ãĢį

-0.66

 Remix

-0.66

 blends

-0.65

POSITIVE LOGITS

DOJ

1.27

 unconstitutional

1.26

 prosecut

1.25

 impunity

1.23

 perjury

1.23

 taxpayer

1.17

 whistleblowers

1.17

 taxpayers

1.13

 treason

1.12

 lawy

1.11

Activations Density 1.046%

legal and political terms, including words related to ethics and institutions

phrases related to hate speech and its implications

No Comments

No Known Activations