INDEX

Explanations

positive or exciting things described as "awesome" with a strong activation value

oai_token-act-pair · gpt-3.5-turbo

expressions of strong enthusiasm or admiration

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

licts

-0.71

mediate

-0.70

uting

-0.68

avis

-0.67

arenthood

-0.67

vere

-0.66

epad

-0.65

epend

-0.64

unker

-0.64

utions

-0.64

POSITIVE LOGITS

ly

0.99

NESS

0.88

 Awesome

0.79

fun

0.77

 awesome

0.77

 teamwork

0.76

 stuff

0.74

ness

0.73

 sounding

0.72

ery

0.72

Activations Density 0.027%

positive or exciting things described as "awesome" with a strong activation value

expressions of strong enthusiasm or admiration

No Comments

No Known Activations