INDEX

Explanations

references to the Defense Department

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "Defense" related to military or governmental contexts

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

bern

-0.73

otin

-0.72

hell

-0.72

hey

-0.69

olics

-0.67

à

-0.66

hest

-0.66

shot

-0.65

pel

-0.65

jury

-0.64

POSITIVE LOGITS

 Defense

1.00

 contractor

0.98

 Forces

0.91

 contractors

0.84

 Authorization

0.83

 Force

0.83

 Ministry

0.82

 Department

0.82

 Dept

0.81

 Memor

0.80

Activations Density 0.018%

references to the Defense Department

mentions of the term "Defense" related to military or governmental contexts

No Comments

No Known Activations