INDEX

Explanations

words related to legal actions, specifically preemptive actions

oai_token-act-pair · gpt-3.5-turbo

concepts related to preemption and preventative measures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Guy

-0.72

 slides

-0.67

 renowned

-0.66

 Anat

-0.64

Bam

-0.63

 studies

-0.63

 located

-0.63

ä¸Ń

-0.63

wi

-0.62

Vu

-0.62

POSITIVE LOGITS

empt

4.49

emption

2.27

emp

1.24

empl

1.23

eem

1.21

ourney

1.08

EMP

1.06

 preempt

1.01

rompt

0.95

hent

0.92

Activations Density 0.012%

words related to legal actions, specifically preemptive actions

concepts related to preemption and preventative measures

No Comments

No Known Activations