INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names, predominately related to individuals

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

å¥³

-0.69

MAC

-0.66

ulhu

-0.64

raped

-0.64

uously

-0.61

 depend

-0.60

yip

-0.59

AST

-0.59

 FANTASY

-0.56

 Leia

-0.55

POSITIVE LOGITS

uner

0.79

ĸļ

0.73

enhagen

0.73

backer

0.71

runner

0.70

onson

0.70

zen

0.69

roth

0.69

nih

0.68

enberg

0.67

Activations Density 0.132%

names of individuals

proper nouns, particularly names, predominately related to individuals

No Comments

No Known Activations