INDEX

Explanations

words related to theft and property crimes

oai_token-act-pair · gpt-3.5-turbo

words related to burglary and criminal activities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

shr

-0.90

LESS

-0.79

acebook

-0.76

eem

-0.74

utical

-0.74

hair

-0.74

EVA

-0.71

rendered

-0.69

usha

-0.69

OPT

-0.69

POSITIVE LOGITS

 burgl

1.39

 burg

1.38

 burglary

1.25

 alarms

0.85

 robbery

0.85

 spree

0.77

 remod

0.77

 intruder

0.76

 leve

0.75

 wiret

0.75

Activations Density 0.008%

words related to theft and property crimes

words related to burglary and criminal activities

No Comments

No Known Activations