INDEX

Explanations

phrases related to physical attacks or criticisms

oai_token-act-pair · gpt-3.5-turbo

terms related to aggressive or violent actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

soDeliveryDate

-0.76

inder

-0.75

inders

-0.74

BuyableInstoreAndOnline

-0.73

fw

-0.69

hander

-0.67

shown

-0.66

vent

-0.66

âĢ¢âĢ¢âĢ¢âĢ¢

-0.66

isphere

-0.65

POSITIVE LOGITS

 attacks

0.75

 against

0.73

ively

0.71

 attackers

0.67

 spree

0.66

 attacking

0.65

ivated

0.64

 attack

0.64

 creeps

0.63

 inflicting

0.63

Activations Density 0.031%

phrases related to physical attacks or criticisms

terms related to aggressive or violent actions

No Comments

No Known Activations