INDEX

Explanations

mentions of military entities, specifically focusing on the "Joint Chiefs of Staff."

oai_token-act-pair · gpt-3.5-turbo

references to military or governmental committees and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Giul

-0.70

cul

-0.69

haps

-0.66

utics

-0.66

 beware

-0.66

acious

-0.66

rx

-0.66

guns

-0.64

+++

-0.64

xual

-0.64

POSITIVE LOGITS

 Chiefs

0.85

 Forces

0.83

 Junction

0.81

 Resolution

0.79

ness

0.77

 Force

0.76

0.74

 Venture

0.73

TAIN

0.72

 Strike

0.70

Activations Density 0.023%

mentions of military entities, specifically focusing on the "Joint Chiefs of Staff."

references to military or governmental committees and organizations

No Comments

No Known Activations