INDEX

Explanations

phrases related to specific events or occurrences, such as injuries, arrests, or evictions

oai_token-act-pair · gpt-3.5-turbo

references to incidents involving harm or injury

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

iqueness

-0.61

polit

-0.60

Therefore

-0.58

Definition

-0.55

inal

-0.55

 insofar

-0.53

 domestically

-0.53

milo

-0.52

Category

-0.52

Â¢

-0.51

POSITIVE LOGITS

tower

0.49

 Survivors

0.48

 pics

0.47

HOU

0.46

 unsuspecting

0.46

 festive

0.46

Sew

0.46

 barric

0.46

 convoy

0.45

 COURT

0.45

Activations Density 1.698%

phrases related to specific events or occurrences, such as injuries, arrests, or evictions

references to incidents involving harm or injury

No Comments

No Known Activations