INDEX

Explanations

verbs related to making mistakes and causing problems for oneself

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

expressions related to mistakes or disruptions

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

vation

-0.85

Interstitial

-0.75

uti

-0.71

eele

-0.62

 Citation

-0.61

baugh

-0.61

åī

-0.61

icular

-0.61

Footnote

-0.58

neum

-0.57

POSITIVE LOGITS

around

1.18

 around

1.16

 Around

0.98

Around

0.93

 havoc

0.91

drivers

0.86

driver

0.81

amn

0.75

 things

0.75

ily

0.72

Activations Density 0.146%

verbs related to making mistakes and causing problems for oneself

expressions related to mistakes or disruptions

No Comments

No Known Activations