INDEX

Explanations

leadership positions such as "Chair" or specific committees

oai_token-act-pair · gpt-3.5-turbo

phrases indicating leadership roles or positions of authority

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

URA

-0.74

eni

-0.71

 experien

-0.65

Db

-0.65

QL

-0.65

PDATE

-0.63

 abund

-0.62

ONSORED

-0.62

aris

-0.61

anmar

-0.61

POSITIVE LOGITS

ĵĺ

0.80

staff

0.75

Roll

0.72

itone

0.71

ãģ®å®

0.69

OTUS

0.69

Klu

0.67

pole

0.67

 desk

0.66

lder

0.66

Activations Density 0.215%

leadership positions such as "Chair" or specific committees

phrases indicating leadership roles or positions of authority

No Comments

No Known Activations