INDEX

Explanations

terms related to physical harm or destruction

oai_token-act-pair · gpt-3.5-turbo

references to physical or structural damage

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

KER

-0.68

ulated

-0.67

arte

-0.67

rams

-0.65

zee

-0.64

 Kong

-0.63

ulates

-0.62

ILLE

-0.61

ramid

-0.61

 Helpful

-0.60

POSITIVE LOGITS

 damage

1.21

 inflicted

1.05

damage

0.97

 mitigation

0.89

 damaged

0.86

 damages

0.81

 Damage

0.78

 horm

0.77

Damage

0.77

 undermin

0.77

Activations Density 0.012%

terms related to physical harm or destruction

references to physical or structural damage

No Comments

No Known Activations