INDEX

Explanations

names of individuals in a structured format, potentially linking them to specific activities or events

oai_token-act-pair · gpt-3.5-turbo

names of individuals and characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

¶ħ

-0.68

 Flavoring

-0.67

berra

-0.62

cffff

-0.58

 âĢº

-0.56

taboola

-0.55

ymm

-0.54

 Sabha

-0.52

AMD

-0.52

ËĪ

-0.52

POSITIVE LOGITS

 disqualified

0.61

's

0.59

ey

0.59

 vetoed

0.58

 waived

0.58

enson

0.57

 accuser

0.57

igham

0.55

failed

0.55

kson

0.54

Activations Density 0.777%

names of individuals in a structured format, potentially linking them to specific activities or events

names of individuals and characters

No Comments

No Known Activations