INDEX

Explanations

phrases related to criminal investigations and law enforcement

oai_token-act-pair · gpt-3.5-turbo

negative indicators or terms associated with investigations and legal matters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

htt

-0.70

 batting

-0.67

cir

-0.66

 caution

-0.66

ieri

-0.66

 Prometheus

-0.64

 context

-0.64

 bounce

-0.63

zip

-0.63

 Norris

-0.63

POSITIVE LOGITS

mediated

1.21

sama

1.17

derived

1.13

induced

1.06

style

1.05

cert

1.04

related

1.02

MAG

0.99

based

0.99

driven

0.99

Activations Density 0.119%

phrases related to criminal investigations and law enforcement

negative indicators or terms associated with investigations and legal matters

No Comments

No Known Activations