INDEX

Explanations

the word "department"

oai_token-act-pair · gpt-3.5-turbo

references to various departments in a context related to institutions or organizations

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

telling

-0.78

 Intent

-0.73

isers

-0.67

Mini

-0.65

iser

-0.64

lihood

-0.60

pak

-0.59

 Instruments

-0.59

chers

-0.59

isks

-0.59

POSITIVE LOGITS

artment

0.97

al

0.92

alities

0.90

artments

0.89

utical

0.86

alse

0.85

ĳå£«

0.83

naire

0.82

ality

0.82

istry

0.81

Activations Density 0.020%

the word "department"

references to various departments in a context related to institutions or organizations

No Comments

No Known Activations