INDEX

Explanations

words related to mechanisms or inner workings

oai_token-act-pair · gpt-3.5-turbo

references to the inner workings and contents of various systems or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ALLY

-0.76

\\\\\\\\

-0.70

oni

-0.69

asus

-0.63

ANN

-0.62

rone

-0.61

IFIED

-0.60

umsy

-0.59

IQ

-0.59

ulus

-0.58

POSITIVE LOGITS

 thereof

1.35

pring

1.24

of

1.21

cape

1.07

creen

1.02

hift

1.00

pire

1.00

pace

0.99

mith

0.96

hip

0.92

Activations Density 0.334%

words related to mechanisms or inner workings

references to the inner workings and contents of various systems or situations

No Comments

No Known Activations