INDEX

Explanations

proper nouns such as names of cities, titles of works, and places

oai_token-act-pair · gpt-3.5-turbo

titles and names of films or media entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 confir

-0.89

actionDate

-0.86

 sanction

-0.84

etheless

-0.80

itionally

-0.79

 reputable

-0.78

 estimating

-0.76

 streng

-0.76

umerous

-0.74

 cryptoc

-0.74

POSITIVE LOGITS

 Revis

1.44

 Presents

1.39

 Massacre

1.33

Pt

1.28

 Orchestra

1.28

 Memories

1.22

 Redux

1.19

 Begins

1.18

 Trilogy

1.15

 Featuring

1.14

Activations Density 0.282%

proper nouns such as names of cities, titles of works, and places

titles and names of films or media entities

No Comments

No Known Activations