INDEX

Explanations

terms related to negative actions or outcomes

oai_token-act-pair · gpt-3.5-turbo

terms related to economic and political issues

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

ortium

-0.67

ulhu

-0.65

ovie

-0.62

undrum

-0.62

eanor

-0.61

DragonMagazine

-0.60

arnaev

-0.60

ilogy

-0.60

utra

-0.59

ndum

-0.59

POSITIVE LOGITS

bugs

0.64

 warfare

0.58

 margins

0.57

 purposes

0.57

 levels

0.56

 tissue

0.55

 situations

0.54

 emergencies

0.53

 farming

0.51

 fatigue

0.51

Activations Density 1.029%

terms related to negative actions or outcomes

terms related to economic and political issues

No Comments

No Known Activations