INDEX

Explanations

phrases related to discrediting or intimidating actions

oai_token-act-pair · gpt-3.5-turbo

phrases related to discrediting and undermining individuals or groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Borders

-0.68

 Rolls

-0.68

Tok

-0.67

 Skies

-0.67

 Jets

-0.67

 Dungeons

-0.67

 Colts

-0.66

 Horse

-0.66

 Shots

-0.65

 Animation

-0.65

POSITIVE LOGITS

rogen

1.12

rogens

1.06

igmat

0.92

 dissemin

0.90

 rehabilit

0.88

 amplify

0.87

 eliminate

0.83

 analyse

0.79

 humili

0.79

 defend

0.79

Activations Density 0.141%

phrases related to discrediting or intimidating actions

phrases related to discrediting and undermining individuals or groups

No Comments

No Known Activations