INDEX

Explanations

hyperlinks indicated by the word "here" for users to click on

oai_token-act-pair · gpt-3.5-turbo

hyperlinks or calls to action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

anking

-0.67

cade

-0.63

Edge

-0.63

ushed

-0.57

tnc

-0.55

luaj

-0.55

opoulos

-0.53

ourn

-0.53

awed

-0.53

Introduced

-0.53

POSITIVE LOGITS

 Oops

0.91

tical

0.78

abouts

0.75

 LINK

0.74

 Attribution

0.73

tics

0.73

>>

0.72

 NEXT

0.72

for

0.71

itness

0.67

Activations Density 0.023%

hyperlinks indicated by the word "here" for users to click on

hyperlinks or calls to action

No Comments

No Known Activations