INDEX

Explanations

comparative phrases indicating increasing intensity or correlation between actions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating increasing intensity or complexity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

é¾įå

-0.87

çīĪ

-0.76

poke

-0.73

oths

-0.69

ospons

-0.66

unic

-0.66

rities

-0.64

esides

-0.64

IVERS

-0.62

 john

-0.62

POSITIVE LOGITS

 chance

0.84

 chances

0.76

 likely

0.73

 tendency

0.70

 likelihood

0.68

raq

0.68

 probability

0.67

CHA

0.66

 temptation

0.65

 havoc

0.62

Activations Density 0.062%

comparative phrases indicating increasing intensity or correlation between actions

phrases indicating increasing intensity or complexity

No Comments

No Known Activations