INDEX

Explanations

words related to burglary and criminal activity

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.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

nesota

-0.84

itism

-0.75

Reviewer

-0.74

 Nadu

-0.70

 Shutterstock

-0.66

 Mara

-0.61

manship

-0.59

ORTS

-0.59

iasm

-0.59

 lear

-0.58

POSITIVE LOGITS

undy

1.20

lar

1.02

dor

0.96

geoning

0.94

 burg

0.86

tone

0.85

untu

0.79

phen

0.77

acies

0.76

auld

0.76

Activations Density 0.024%

words related to burglary and criminal activity

No Comments

No Known Activations