INDEX

Explanations

phrases related to news headlines or important events

oai_token-act-pair · gpt-3.5-turbo

repetitive calls to action for replaying content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

pora

-0.73

zbek

-0.72

kers

-0.71

EStreamFrame

-0.71

krit

-0.70

bara

-0.68

ocene

-0.66

ksh

-0.65

 sucker

-0.64

esan

-0.63

POSITIVE LOGITS

 Replay

0.93

 WATCHED

0.77

lements

0.73

INGS

0.73

 Thumbnails

0.73

iece

0.71

Ħ¢

0.68

artz

0.68

 Videos

0.67

actions

0.66

Activations Density 0.011%

phrases related to news headlines or important events

repetitive calls to action for replaying content

No Comments

No Known Activations