INDEX

Explanations

phrases indicating understanding, justification, or explanation

oai_token-act-pair · gpt-3.5-turbo

phrases indicating comprehension or reasonableness of actions

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

etry

-0.73

esh

-0.70

eki

-0.69

pee

-0.67

eng

-0.66

infect

-0.66

ngth

-0.65

andon

-0.63

 resh

-0.62

cler

-0.62

POSITIVE LOGITS

 Mellon

0.83

 indignation

0.80

DragonMagazine

0.75

FontSize

0.74

NPR

0.74

cffffcc

0.72

 understandable

0.70

 outrage

0.68

>>\

0.68

Â¶

0.67

Activations Density 0.048%

phrases indicating understanding, justification, or explanation

phrases indicating comprehension or reasonableness of actions

No Comments

No Known Activations