INDEX

Explanations

instances where procedures, policies, or instructions are emphasized or referenced

oai_token-act-pair · gpt-3.5-turbo

instances of the word "in" and its context within sentences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 summed

-0.66

 laughs

-0.64

%%

-0.62

 pins

-0.61

 grows

-0.60

CHA

-0.59

âĿ

-0.59

 fame

-0.59

 classmates

-0.59

 alas

-0.58

POSITIVE LOGITS

 lieu

1.48

 accordance

1.45

 conjunction

1.33

 relation

1.22

appropriate

1.16

ordinate

1.15

 regards

1.13

effic

1.11

 favour

1.10

 favor

1.10

Activations Density 0.400%

instances where procedures, policies, or instructions are emphasized or referenced

instances of the word "in" and its context within sentences

No Comments

No Known Activations