INDEX

Explanations

specific instances of people joining something new or a group

oai_token-act-pair · gpt-3.5-turbo

instances of the word "joined" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Merry

-0.68

binary

-0.65

ester

-0.64

 tuber

-0.64

 tiss

-0.63

framework

-0.63

bane

-0.62

otropic

-0.62

relative

-0.61

andi

-0.61

POSITIVE LOGITS

 join

0.79

ATURES

0.77

Join

0.73

 Join

0.71

join

0.70

ãĥ³

0.69

lehem

0.69

lift

0.68

 joins

0.68

 joining

0.68

Activations Density 0.021%

specific instances of people joining something new or a group

instances of the word "joined" in various contexts

No Comments

No Known Activations