INDEX

Explanations

phrases related to revealing or uncovering something

oai_token-act-pair · gpt-3.5-turbo

references to curtains and related concepts, symbolizing concealment or transition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

versions

-0.70

ctic

-0.70

ishops

-0.69

ciating

-0.69

etically

-0.64

OTAL

-0.63

ogh

-0.63

ese

-0.62

nia

-0.62

omal

-0.61

POSITIVE LOGITS

 curtain

1.34

 curtains

1.20

glass

0.96

 veil

0.83

AGE

0.78

ãĤ¦ãĤ¹

0.77

 Veil

0.75

bridge

0.71

ages

0.69

wat

0.67

Activations Density 0.009%

phrases related to revealing or uncovering something

references to curtains and related concepts, symbolizing concealment or transition

No Comments

No Known Activations