INDEX

Explanations

proper nouns beginning with "Ale" and a high level of activation is seen with words closely related to this proper noun

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Ale" 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

enegger

-0.97

ãĤ¼ãĤ¦ãĤ¹

-0.91

ä¹ĭ

-0.77

ãģį

-0.73

ãģı

-0.71

 Everyday

-0.71

CFR

-0.67

aneously

-0.67

ledged

-0.64

 prominently

-0.63

POSITIVE LOGITS

jandro

1.21

ppo

1.02

lette

0.94

oga

0.93

chemy

0.88

una

0.87

terness

0.86

Ale

0.85

rique

0.82

bum

0.81

Activations Density 0.012%

proper nouns beginning with "Ale" and a high level of activation is seen with words closely related to this proper noun

mentions of the word "Ale" in various contexts

No Comments

No Known Activations