INDEX

Explanations

statements where someone is mentioned to have not spoken or said anything in a given situation

oai_token-act-pair · gpt-3.5-turbo

instances of speech or communication about significant issues

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ymph

-0.63

eele

-0.63

acha

-0.63

ipel

-0.62

omorph

-0.60

ortment

-0.60

vy

-0.60

inen

-0.60

unning

-0.58

 Tycoon

-0.57

POSITIVE LOGITS

 whatsoever

1.30

nor

1.18

 except

1.15

 until

1.14

 anymore

1.13

 till

1.11

until

1.06

 unless

0.93

 bothered

0.85

 anywhere

0.82

Activations Density 0.487%

statements where someone is mentioned to have not spoken or said anything in a given situation

instances of speech or communication about significant issues

No Comments

No Known Activations