INDEX

Explanations

phrases indicating awareness or curiosity about ongoing situations or events

oai_token-act-pair · gpt-3.5-turbo

phrases that inquire about ongoing events or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

¿½

-0.90

ĳå£«

-0.79

ritical

-0.74

ugi

-0.73

ĵ

-0.71

vale

-0.67

guyen

-0.65

ĭ

-0.64

ForgeModLoader

-0.63

ãĥ¼ãĥ³

-0.61

POSITIVE LOGITS

 behind

1.19

 inside

1.16

 backstage

1.07

behind

1.07

 underneath

1.06

 beneath

1.02

 here

1.01

 elsewhere

0.98

 around

0.98

 upstairs

0.97

Activations Density 0.094%

phrases indicating awareness or curiosity about ongoing situations or events

phrases that inquire about ongoing events or situations

No Comments

No Known Activations