INDEX

Explanations

Words related to scientific research and materials, specifically emphasizing the word "material"

oai_token-act-pair · gpt-3.5-turbo

references to the term "ther" 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ELS

-0.87

URI

-0.74

ured

-0.66

 Strauss

-0.64

 Chao

-0.64

 Lomb

-0.63

 Peel

-0.63

 NYPD

-0.62

 admirable

-0.61

POSITIVE LOGITS

apy

1.41

idge

1.14

apist

1.12

swer

1.07

netic

1.06

ther

1.03

nian

0.99

soever

0.90

iated

0.89

ington

0.89

Activations Density 0.004%

Words related to scientific research and materials, specifically emphasizing the word "material"

references to the term "ther" in various contexts

No Comments

No Known Activations