INDEX

Explanations

adjectives describing characteristics or qualities

oai_token-act-pair · gpt-3.5-turbo

various negative or critical descriptors and themes

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

DonaldTrump

-0.67

iatus

-0.65

outube

-0.62

ilaterally

-0.61

itored

-0.61

;;;;

-0.61

DoS

-0.61

terday

-0.60

utics

-0.60

ecause

-0.60

POSITIVE LOGITS

iest

0.95

 equivalent

0.79

liest

0.79

 element

0.73

 closest

0.72

 portion

0.71

 nearest

0.70

ultimate

0.68

 itself

0.66

 version

0.66

Activations Density 0.906%

adjectives describing characteristics or qualities

various negative or critical descriptors and themes

No Comments

No Known Activations