INDEX

Explanations

phrases related to individuals joining specific groups or events

oai_token-act-pair · gpt-3.5-turbo

instances of the word "joined."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

binary

-0.68

ester

-0.66

relative

-0.64

inarily

-0.63

bane

-0.62

 tuber

-0.62

framework

-0.61

aha

-0.61

otropic

-0.61

 Merry

-0.60

POSITIVE LOGITS

âĸ¬âĸ¬

0.71

 Join

0.70

Join

0.70

 join

0.68

jen

0.68

ãĤ¤ãĥĪ

0.67

join

0.67

 joins

0.67

 joined

0.66

ATURES

0.66

Activations Density 0.018%

phrases related to individuals joining specific groups or events

instances of the word "joined."

No Comments

No Known Activations