INDEX

Explanations

phrases highlighting the uniqueness or novelty of something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating unprecedented experiences or situations

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

Talk

-0.66

abet

-0.65

catentry

-0.64

Michelle

-0.63

bern

-0.63

Activ

-0.62

phone

-0.62

Alexander

-0.61

Occup

-0.61

raz

-0.61

POSITIVE LOGITS

cies

0.79

aneously

0.74

afort

0.70

orthy

0.69

htaking

0.69

 realised

0.68

 fading

0.68

 existed

0.67

 tasted

0.67

 realizing

0.66

Activations Density 0.038%

phrases highlighting the uniqueness or novelty of something

phrases indicating unprecedented experiences or situations

No Comments

No Known Activations