INDEX

Explanations

words related to specific technologies or technical terms

oai_token-act-pair · gpt-3.5-turbo

elements related to film or cinematic topics

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

 reconc

-0.63

 perspect

-0.60

 tremend

-0.60

 undermin

-0.59

 horizont

-0.59

 simultane

-0.55

 Palestin

-0.54

depending

-0.53

 stitching

-0.52

 neighb

-0.51

POSITIVE LOGITS

 âĵĺ

0.65

eus

0.62

ensis

0.55

gmail

0.53

reme

0.52

ouse

0.49

ride

0.48

otto

0.47

 Rally

0.47

ukong

0.47

Activations Density 0.910%

words related to specific technologies or technical terms

elements related to film or cinematic topics

No Comments

No Known Activations