INDEX

Explanations

phrases describing people, their attributes, or their relationships with others

oai_token-act-pair · gpt-3.5-turbo

references to individuals identified by the word "whose."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

hal

-0.71

ï¿½ï¿½ï¿½ï¿½

-0.70

enge

-0.70

yi

-0.68

edd

-0.68

CLASS

-0.67

hari

-0.66

pn

-0.65

yang

-0.65

rator

-0.65

POSITIVE LOGITS

 sole

1.11

 ancestors

1.06

own

1.04

 whereabouts

0.91

 estates

0.87

 fault

0.84

 entire

0.83

 opinions

0.83

 names

0.82

 namesake

0.82

Activations Density 0.031%

phrases describing people, their attributes, or their relationships with others

references to individuals identified by the word "whose."

No Comments

No Known Activations