INDEX

Explanations

adjectives related to harm or negative impact

oai_token-act-pair · gpt-3.5-turbo

terms related to destruction and harmful effects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Cosponsors

-0.76

ask

-0.74

Clar

-0.72

shire

-0.71

HR

-0.66

 Patel

-0.66

icip

-0.66

 âī¤

-0.64

 Lilly

-0.64

 Kislyak

-0.64

POSITIVE LOGITS

 destructive

3.19

destruct

2.38

 destruct

2.01

 destroying

1.88

 destruction

1.79

dest

1.67

Dest

1.66

 Destruction

1.64

 Dest

1.55

 sabot

1.44

Activations Density 0.032%

adjectives related to harm or negative impact

terms related to destruction and harmful effects

No Comments

No Known Activations