INDEX

Explanations

references to the name "Yoda" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to specific names and terms associated with individuals or characters

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

dar

-0.72

creen

-0.71

line

-0.69

 Chilean

-0.67

hips

-0.66

 Agency

-0.65

mut

-0.65

reach

-0.65

ety

-0.65

ray

-0.64

POSITIVE LOGITS

uyomi

0.85

xual

0.85

pmwiki

0.82

ajo

0.82

utra

0.82

itsch

0.80

Mn

0.79

plin

0.77

enegger

0.77

aceutical

0.74

Activations Density 0.064%

references to the name "Yoda" in various contexts

references to specific names and terms associated with individuals or characters

No Comments

No Known Activations