INDEX

Explanations

information related to key events, developments, and actions

oai_token-act-pair · gpt-3.5-turbo

action words associated with initiating or documenting significant events or processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

hart

-0.72

ennett

-0.70

-0.65

antics

-0.64

 Gentleman

-0.64

mania

-0.63

 coun

-0.61

Pak

-0.61

xus

-0.61

nir

-0.60

POSITIVE LOGITS

 originally

0.71

 miscar

0.70

 previously

0.67

 Lerner

0.67

 fasc

0.66

 formerly

0.66

 Yanuk

0.64

 hitherto

0.62

 Tyrann

0.60

é¾įå¥ĳå£«

0.60

Activations Density 0.243%

information related to key events, developments, and actions

action words associated with initiating or documenting significant events or processes

No Comments

No Known Activations