INDEX

Explanations

words related to legal or criminal activities, specifically focusing on unlawful actions

oai_token-act-pair · gpt-3.5-turbo

references to legal terms regarding unlawful and lawful actions

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

 Solitaire

-0.81

 Tycoon

-0.78

bol

-0.77

chell

-0.73

olog

-0.73

ocl

-0.71

acea

-0.70

 Atlas

-0.69

oche

-0.67

lasses

-0.67

POSITIVE LOGITS

theless

1.01

00200000

0.87

cy

0.80

isance

0.76

 possession

0.75

 unfocusedRange

0.75

ifiable

0.75

SPONSORED

0.74

rontal

0.72

 ACTIONS

0.70

Activations Density 0.022%

words related to legal or criminal activities, specifically focusing on unlawful actions

references to legal terms regarding unlawful and lawful actions

No Comments

No Known Activations