INDEX

Explanations

variations of the word "sick" with increasing intensity

oai_token-act-pair · gpt-3.5-turbo

repeated and excessive use of the word "sick" 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 @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ĵĺ

-0.82

CLASSIFIED

-0.62

wcs

-0.60

 Corpus

-0.57

 Adren

-0.56

 festive

-0.52

Kick

-0.51

 Brist

-0.51

 Medals

-0.49

 Thanksgiving

-0.49

POSITIVE LOGITS

onga

0.61

orage

0.60

osuke

0.60

rafted

0.57

eto

0.57

lings

0.54

miah

0.54

obyl

0.54

 hers

0.54

rite

0.53

Activations Density 4.072%

variations of the word "sick" with increasing intensity

repeated and excessive use of the word "sick" in various contexts

No Comments

No Known Activations