INDEX

Explanations

proper nouns related to law enforcement agencies or individuals, particularly those related to Broward County

oai_token-act-pair · gpt-3.5-turbo

references to Broward County

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ãĥ¤

-0.86

 Haram

-0.82

 Romeo

-0.75

 Jordanian

-0.68

ãĥ¼ãĥ³

-0.68

BMC

-0.66

 Hungarian

-0.66

 pastoral

-0.65

ãĤª

-0.65

 Amos

-0.64

POSITIVE LOGITS

 Brow

1.23

Brow

1.09

atche

0.98

byss

0.97

der

0.89

count

0.86

halla

0.85

swer

0.85

pedia

0.84

beat

0.84

Activations Density 0.013%

proper nouns related to law enforcement agencies or individuals, particularly those related to Broward County

references to Broward County

No Comments

No Known Activations