INDEX

Explanations

locations and geographical features

oai_token-act-pair · gpt-3.5-turbo

references to areas impacted by conflict or war

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

 Lifetime

-0.69

 WARRANT

-0.55

ester

-0.55

Reference

-0.54

 msec

-0.54

Act

-0.54

BAT

-0.54

 pedals

-0.53

aughs

-0.53

redo

-0.52

POSITIVE LOGITS

 inhabited

0.92

 UNESCO

0.90

 adjoining

0.87

 populated

0.83

 inhabitants

0.82

 landmarks

0.79

 adjacent

0.77

Located

0.77

 wetlands

0.77

 bustling

0.76

Activations Density 1.101%

locations and geographical features

references to areas impacted by conflict or war

No Comments

No Known Activations