INDEX

Explanations

the word "that" at a high activation value

oai_token-act-pair · gpt-3.5-turbo

deeply embedded phrases that emphasize the concept of "that"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

istance

-0.71

notice

-0.68

ãĥ©ãĥ³

-0.68

hips

-0.62

ãĥīãĥ©

-0.62

thro

-0.62

 Directions

-0.60

idelines

-0.60

Guard

-0.58

ãĥĺ

-0.58

POSITIVE LOGITS

cher

0.85

 justifies

0.74

 contradicts

0.67

 translates

0.66

 sounded

0.63

 kind

0.63

 sounds

0.63

 qualifies

0.62

 fits

0.62

 applies

0.61

Activations Density 0.296%

the word "that" at a high activation value

deeply embedded phrases that emphasize the concept of "that"

No Comments

No Known Activations