INDEX

Explanations

references to specific locations in conflict zones

oai_token-act-pair · gpt-3.5-turbo

references to locations involved in conflict

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ype

-0.72

olson

-0.71

writers

-0.70

morph

-0.68

ISON

-0.67

ãĤ©

-0.67

arious

-0.67

 Nguyen

-0.66

pn

-0.65

milo

-0.64

POSITIVE LOGITS

 Province

1.26

 province

1.19

 countryside

1.12

 neighbourhoods

1.01

 stronghold

0.98

 municipality

0.96

 Heights

0.95

 city

0.93

 Municip

0.93

 neighborhoods

0.93

Activations Density 0.100%

references to specific locations in conflict zones

references to locations involved in conflict

No Comments

No Known Activations