INDEX

Explanations

references to specific companies or organizations as well as words related to negative social behaviors

oai_token-act-pair · gpt-3.5-turbo

references to companies and antisemitism

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Nile

-0.66

åĮ

-0.66

tails

-0.65

isen

-0.63

eering

-0.61

ogy

-0.60

 heads

-0.59

 Worlds

-0.59

itarian

-0.58

arty

-0.58

POSITIVE LOGITS

aurus

1.25

creen

1.23

ystem

1.20

earch

1.17

erver

1.16

ocial

1.12

ullivan

1.10

hiba

1.09

cript

1.05

omething

1.04

Activations Density 0.057%

references to specific companies or organizations as well as words related to negative social behaviors

references to companies and antisemitism

No Comments

No Known Activations