INDEX

Explanations

the word "ricks" with a varying level of intensity

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of a specific individual's last name

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Monitor

-0.86

CONT

-0.73

 transmitted

-0.67

 Palest

-0.65

 unfavorable

-0.65

 reproduction

-0.65

 archived

-0.63

 Administrator

-0.63

 Pandora

-0.63

ãĥ´

-0.62

POSITIVE LOGITS

ricks

4.46

riks

2.64

kson

1.23

rists

1.19

dies

1.18

iots

1.17

atars

1.15

eways

1.13

ifts

1.11

riots

1.11

Activations Density 0.022%

the word "ricks" with a varying level of intensity

repeated mentions of a specific individual's last name

No Comments

No Known Activations