INDEX

Explanations

phrases related to fullness or completeness

oai_token-act-pair · gpt-3.5-turbo

references to completeness or wholeness across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

Phi

-0.71

 pige

-0.69

 Morse

-0.66

Palest

-0.62

chance

-0.62

zees

-0.60

apest

-0.60

ammers

-0.59

gat

-0.59

bably

-0.59

POSITIVE LOGITS

ledged

0.90

fledged

0.73

Initialized

0.71

âĶĢâĶĢâĶĢâĶĢ

0.67

insula

0.64

urse

0.63

ACTION

0.63

uberty

0.62

pection

0.62

ngth

0.61

Activations Density 0.220%

phrases related to fullness or completeness

references to completeness or wholeness across various contexts

No Comments

No Known Activations