INDEX

Explanations

words related to legal terms and actions

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to large quantities or amounts related to spending or consumption

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zag

-0.74

cent

-0.63

opter

-0.62

dfx

-0.62

ocamp

-0.61

insula

-0.61

ufact

-0.61

soType

-0.59

yrinth

-0.58

animous

-0.56

POSITIVE LOGITS

 discrepancies

0.61

raits

0.59

>[

0.59

 Views

0.58

aways

0.55

 aback

0.55

 rights

0.54

 disputed

0.53

0.52

 IMAGES

0.51

Activations Density 0.258%

words related to legal terms and actions

references to large quantities or amounts related to spending or consumption

No Comments

No Known Activations