INDEX

Explanations

references to violent actions with high casualty rates

oai_token-act-pair · gpt-3.5-turbo

references to casualties and deadly incidents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

itus

-0.74

gaard

-0.73

chart

-0.69

 athlet

-0.68

anwhile

-0.68

stage

-0.67

uyomi

-0.66

jri

-0.65

ãĤ´ãĥ³

-0.65

 occas

-0.63

POSITIVE LOGITS

him

1.14

 civilians

0.86

 them

0.84

 anyone

0.80

 anybody

0.76

 somebody

0.76

 whoever

0.75

 someone

0.75

 foreigners

0.74

 informants

0.74

Activations Density 0.183%

references to violent actions with high casualty rates

references to casualties and deadly incidents

No Comments

No Known Activations