INDEX

Explanations

specific terms related to legality or regulations, particularly those containing the substring "pro"

oai_token-act-pair · gpt-3.5-turbo

the prefix "pro," indicating support for various progressive or pro-choice concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ãĤ¼ãĤ¦ãĤ¹

-0.81

 needles

-0.77

 Twain

-0.73

 Halls

-0.72

 bedrooms

-0.72

 spears

-0.69

 strawberries

-0.67

hyde

-0.66

inia

-0.66

ĸļ

-0.66

POSITIVE LOGITS

digy

1.28

blems

1.27

secut

1.26

ceed

1.24

secution

1.17

gressive

1.11

posal

1.10

bable

1.08

gression

1.06

cess

1.01

Activations Density 0.020%

specific terms related to legality or regulations, particularly those containing the substring "pro"

the prefix "pro," indicating support for various progressive or pro-choice concepts

No Comments

No Known Activations