INDEX

Explanations

words related to criminal activity or negative events

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "ab" in various contexts

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

 nomine

-0.84

 wills

-0.76

 virtue

-0.71

 sclerosis

-0.69

backer

-0.69

bilt

-0.68

 Honour

-0.67

 withdrawals

-0.66

 Titanic

-0.65

 Maker

-0.64

POSITIVE LOGITS

stract

1.33

raham

1.11

yrinth

1.10

bing

1.07

ril

1.00

ras

0.99

uilt

0.99

road

0.99

iotics

0.99

bed

0.98

Activations Density 0.023%

words related to criminal activity or negative events

occurrences of the substring "ab" in various contexts

No Comments

No Known Activations