INDEX

Explanations

instances where one thing is compared or contrasted with another using the word "opposed"

oai_token-act-pair · gpt-3.5-turbo

comparisons involving the phrase "as opposed to."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

redo

-0.75

 Highlights

-0.65

eda

-0.63

 Ceres

-0.60

 Conquer

-0.58

 Naples

-0.57

 trave

-0.57

 Dahl

-0.57

 nodd

-0.55

 Verd

-0.55

POSITIVE LOGITS

 thereto

1.01

to

0.88

thodox

0.68

arth

0.68

hammad

0.67

 merely

0.66

ot

0.66

 "$:/

0.64

uitive

0.63

acles

0.63

Activations Density 0.032%

instances where one thing is compared or contrasted with another using the word "opposed"

comparisons involving the phrase "as opposed to."

No Comments

No Known Activations