INDEX

Explanations

instances where something is deemed excessive or surpassing a certain limit

oai_token-act-pair · gpt-3.5-turbo

instances of the word "too" indicating excessive or excessive amounts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

inarily

-0.67

è¦ļéĨĴ

-0.66

inated

-0.62

Qaeda

-0.61

makes

-0.60

inates

-0.60

oust

-0.60

riots

-0.60

tu

-0.60

 Avalanche

-0.59

POSITIVE LOGITS

imei

0.73

OHN

0.68

 numb

0.65

ħĭ

0.65

 bulky

0.65

ĵĺ

0.65

akening

0.65

raviolet

0.63

 fatty

0.63

 much

0.63

Activations Density 0.033%

instances where something is deemed excessive or surpassing a certain limit

instances of the word "too" indicating excessive or excessive amounts

No Comments

No Known Activations