INDEX

Explanations

positive adjectives, particularly the word "good"

oai_token-act-pair · gpt-3.5-turbo

the word "good" and its variations, indicating positive assessments or qualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Enlarge

-0.78

heim

-0.69

nai

-0.69

stadt

-0.68

rop

-0.67

ansk

-0.66

igans

-0.65

igate

-0.64

skirts

-0.63

lander

-0.63

POSITIVE LOGITS

 good

3.11

good

2.37

 GOOD

2.19

 Good

1.91

Good

1.85

 decent

1.84

bad

1.73

 nice

1.72

 excellent

1.67

 great

1.64

Activations Density 0.045%

positive adjectives, particularly the word "good"

the word "good" and its variations, indicating positive assessments or qualities

No Comments

No Known Activations