INDEX

Explanations

references to specific religious groups and financial terms related to payments

oai_token-act-pair · gpt-3.5-turbo

references to religious organizations, specifically Jehovah's Witnesses and the Salvation Army

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

illac

-0.90

iment

-0.73

hips

-0.72

rams

-0.69

mint

-0.68

hip

-0.67

iments

-0.66

etter

-0.65

imental

-0.65

 foliage

-0.64

POSITIVE LOGITS

BOOK

0.70

Universal

0.68

 livelihood

0.67

heit

0.67

ortun

0.66

 fulfillment

0.66

DOM

0.65

Aid

0.65

UTC

0.65

atsu

0.64

Activations Density 0.055%

references to specific religious groups and financial terms related to payments

references to religious organizations, specifically Jehovah's Witnesses and the Salvation Army

No Comments

No Known Activations