INDEX

Explanations

instances of the word "even" followed by a number

oai_token-act-pair · gpt-3.5-turbo

instances of the word "even" used to emphasize comparisons or intensify statements

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rend

-0.87

cel

-0.65

ãĤ¿

-0.64

HCR

-0.63

aim

-0.62

FG

-0.60

reating

-0.60

ync

-0.60

cent

-0.60

onomous

-0.59

POSITIVE LOGITS

 remotely

0.93

 though

0.84

handed

0.74

tho

0.70

ivari

0.67

handedly

0.66

 indirectly

0.60

 moderately

0.59

mont

0.59

 numbered

0.58

Activations Density 0.076%

instances of the word "even" followed by a number

instances of the word "even" used to emphasize comparisons or intensify statements

No Comments

No Known Activations